{
  "startedUtc": "2026-10-06T15:48:37Z",
  "command": "run-all",
  "commandLine": "~/gvb-work/lanes/v7-final/src/GenericVectorBuilder.Bench/bin/Release/net10.0/GenericVectorBuilder.Bench.dll run-all --pipeline eshoponweb --queries golden --seed 703 --out ~/ForClaude/GenericVectorBuilder/bench-results",
  "machine": {
    "host": "bench-host",
    "cpu": "Intel(R) Xeon(R) CPU E5-1620 v3 @ 3.50GHz",
    "logicalCpus": 8,
    "ramGiB": 62.7,
    "gpu": "NVIDIA GeForce RTX 3060, 12288 MiB",
    "os": "Ubuntu 24.04.5 LTS, kernel 6.8.0-142-generic",
    "dotNet": ".NET 10.0.12",
    "loadAverage": "4.99 3.36 3.26"
  },
  "pipeline": "eshoponweb",
  "collection": "gvbbench_eshoponweb",
  "rows": 524,
  "dimension": 1024,
  "source": "Read READ-ONLY from GenericVectorBuilder.dbo.gvb_eshoponweb in ChunkId order, vectors as native SqlVector\u003Cfloat\u003E, 0.2 s.",
  "queries": "golden: 20 labelled questions from ~/ForClaude/evalkit/questions_golden.json, embedded with qwen3-emb-0.6b (cached in ~/gvb-data/bench-cache/golden-eshoponweb-68df42ca69efa079.json, no embedding calls)",
  "queryCount": 20,
  "top": 10,
  "concurrency": [
    1,
    8
  ],
  "secondsPerLevel": 20,
  "truthSeconds": 0.0243631,
  "truthNdcg": 0.5181699774768911,
  "runSeed": 703,
  "targetOrder": [
    "sqlitevec",
    "mariadb",
    "duckdb",
    "qdrant",
    "elasticsearch",
    "vespa",
    "qdrant-hnsw",
    "sql-diskann",
    "pgvector",
    "weaviate",
    "clickhouse",
    "oracle",
    "sql",
    "opensearch",
    "redis",
    "chroma",
    "typesense",
    "mongodb",
    "milvus"
  ],
  "notes": [
    "Order: targets ran one at a time in a random order from seed 703 (runSeed; --seed 703 repeats it, targetOrder lists it). Inside each target the timed passes also ran in a random order from the same seed and the target\u0027s name (passOrder): default@1 is one searcher for 20 s (every search\u0027s latency gives p50/p95/p99, the completed searches give QPS@1, its first answer to each query gives recall and nDCG), default@N is N searchers for 20 s (QPS@N), exact is the engine\u0027s exact mode, one searcher for 60 s cycling the queries.",
    "Preparation, untimed, before any timed pass: a rehearsal of every pass type at its own concurrency for 30 s each, through the same code the passes use.",
    "Warm-up and settle check, untimed, right before every timed pass: the pass\u0027s own search at the pass\u0027s own number of searchers for at least 15 s and at least 20 searches (at most 120 s), read in windows of at least 2 s and 100 searches; then a 3 s trial of the same pass. The trial\u0027s figure (p50 with one searcher, QPS with several) must lie within 10% of the warm-up\u0027s settled figure (the median of its last 3 windows, which must agree within 5%); if not, the warm-up is extended once (at least 30 s, until its windows agree, at most 120 s), where an extension\u0027s windows agree when its level test passes: the older and the newer half of its latest windows, 5 to 10 windows a half, each half read as the pass reads it (the p50 of all its searches with one searcher, its searches per second with several), agree within 5%, judged on the windows that stopped it (an extension whose cap runs out with fewer than 10 windows is judged by its last 3 instead); then a second trial is taken, which must lie inside the range of the newer half\u0027s windows or within 10% of the newer half\u0027s figure, and the warm-up runs again before the pass. After the pass its own figure is held against the settled figure, within 10%. In the level test, the second trial and the hold, two one-searcher p50s no more than 0.1 ms apart agree whatever their percentage (below what this box resolves at one searcher). Each target\u0027s notes give every check, and a pass that still disagrees, or whose own figure did not hold, flags its target as unsettled. With machine control on, the check for a quiet box is made when the warm-up is announced, before it starts (a wait between the warm-up and the timed pass let the engine go cold), so by the time the clock opens that check is as old as the warm-up and the trial (about 18 s, more after an extension); each pass\u0027s conditions record that lead (quietCheckLeadSeconds). Failed untimed searches are counted per target (warmupErrors) and are not in the timed error counts.",
    "Load rows/s counts only time inside each target\u0027s upsert calls: one writer, batches of 1000, rows already in memory, the collection dropped and created fresh first. Engines that build or finish their index after the writes do it in a separate timed index step (the load\u0027s index seconds), and the run waits for it before searching.",
    "Index proof: each engine\u0027s own report of its index (indexState) is read after the load and again after the last pass. A target whose index was not ready after the load was still measured and carries a WARNING; an engine that reports nothing counts as not ready.",
    "Search settings: each searched target records the settings its own index description states (searchSettings: the build parameters and the effort per query, such as m, ef_construction and ef_search), read after the load; consolidating never averages runs whose settings differ.",
    "Durability: each target\u0027s crash-safety setting as configured here (durability). Engines that do not force writes to disk on every commit load faster for that reason.",
    "Latency is client-side wall time around each search (network and driver included), every search of the default@1 window, one searcher, the queries cycled; a window with fewer than 200 searches, a p50 above 1.25 x its mean, or a mean above its p99 is flagged.",
    "QPS: N searchers back to back for 20 s per level; completed searches divided by the window\u0027s elapsed time.",
    "Recall@10: share of the exact top 10 (brute force in memory) that the engine returned. A hit whose exact similarity ties the 10th best (within 1e-5) also counts, because duplicate rows embed to identical vectors.",
    "Routes: a container engine (the benchmark\u0027s own SQL Server and Qdrant containers included) is reached at its container\u0027s own address on its Docker network, never through the published localhost port (docker-proxy); the native comparison targets (sql-native, qdrant-native) are native services reached directly over loopback. Each target\u0027s addresses and the connections the client held open after its passes are in its notes and in conditions.connections.",
    "RAM of compose engines, the benchmark\u0027s own SQL Server and Qdrant containers (sql, sql-diskann, qdrant, qdrant-hnsw) included, is docker stats of the engine\u0027s containers; for the native comparison targets (sql-native, qdrant-native) it is the whole native process, including every other database or collection it serves. Disk is the table\u0027s reserved pages (SQL), the collection folder (Qdrant), or for other engines what the load added to the engine\u0027s data folder (engines that keep data in memory until a snapshot show almost nothing).",
    "Engines: run-all starts an engine that is down and stops it afterwards only if it was not running when the run began; an engine that was already running is left running.",
    "WARNING: the engine did not report a ready index after the load for: elasticsearch. They were measured anyway; their numbers may come from a scan or a half-built index (see each target\u0027s index state).",
    "Machine control on: governor performance on every CPU during the run (before: schedutil; at the end: performance; after putting it back: schedutil). CPU clock pinned for the run: turbo off (intel_pstate/no_turbo 0 -\u003E 1), so every CPU\u0027s clock is held at its ceiling of 3500 MHz whatever the engine runs; the uncore (L3 and memory) clock is held at 3000 MHz (min 3000 MHz, max 3000 MHz (MSR 0x620 = 0x1e1e); was min 1200 MHz, max 3000 MHz (MSR 0x620 = 0xc1e)) (before: turbo on (intel_pstate/no_turbo 0), ceiling 3600 MHz on CPUs 0-7, floor 1200 MHz on CPUs 0-7; while pinned: turbo off (intel_pstate/no_turbo 1), ceiling 3500 MHz on CPUs 0-7, floor 1200 MHz on CPUs 0-7; at the end: turbo off (intel_pstate/no_turbo 1), ceiling 3500 MHz on CPUs 0-7, floor 1200 MHz on CPUs 0-7; after putting it back: turbo on (intel_pstate/no_turbo 0), ceiling 3600 MHz on CPUs 0-7, floor 1200 MHz on CPUs 0-7). Uncore limit at the end: min 3000 MHz, max 3000 MHz (MSR 0x620 = 0x1e1e); after putting it back: min 1200 MHz, max 3000 MHz (MSR 0x620 = 0xc1e). Each target\u0027s clock note gives every pass\u0027s average MHz on the engine CPUs and on the client CPUs; a pass whose average on either is more than 1% off the pinned 3500 MHz is flagged. Figures from runs made with turbo on are not comparable with these in absolute terms. engine CPUs 2-3,6-7 (cores 2,6 and 3,7), client CPUs 0-1,4-5 (cores 0,4 and 1,5); the client process was pinned; each engine was pinned to the engine CPUs for its turn and put back after (conditions.engines); an engine run-all started (and stopped) was asked to be created on them, and its notes say whether the host did so or it was moved there after the start. Busy box: right before each pass\u0027s warm-up the run waits, up to 10 min, while processes outside the benchmark (everything but this client and the engine under test\u0027s cgroups) use more than 0.3 CPUs on average over the last 60 s or the last 5 s, neither window reaching back past the start of the target or of its engine (so the engine\u0027s own start-up is not outside work); if the box does not clear the pass runs anyway, flagged \u0027busy box\u0027, as is a pass whose own outside load is above the limit. Outside load counts busy = user \u002B nice \u002B system \u002B irq \u002B softirq \u002B steal (guest time is already in user); CONFIG_IRQ_TIME_ACCOUNTING is not set, so task and cgroup run time include the interrupt and softirq time that hit them and it is added; CONFIG_PARAVIRT_TIME_ACCOUNTING is not set, so steal is added (/boot/config-6.8.0-142-generic) (conditions.cpuAccounting); each pass also records this client\u0027s own CPU time per search (conditions.passes[].clientCpuMsPerSearch). CPU clocks were sampled every 250 ms; each pass\u0027s min/median/max per CPU is in conditions.passes. CPU idle states, recorded and left as found: driver intel_idle, governor menu, intel_idle max_cstate 9; POLL on, C1 on, C1E on, C3 OFF on CPUs 0-7 (default disabled), C6 on (conditions.cpuIdle). How each target was reached: conditions.connections. Client build Release, .NET 10.0.12.",
    "sqlitevec is embedded: it ran inside the client process on the client CPUs 0-1,4-5, sharing them with the client.",
    "duckdb is embedded: it ran inside the client process on the client CPUs 0-1,4-5, sharing them with the client.",
    "WARNING: busy box during mongodb default@1: processes outside the benchmark used 0.35 CPUs on average during the pass, above the limit of 0.3.",
    "WARNING: busy box during mongodb exact: processes outside the benchmark used 0.30 CPUs on average during the pass, above the limit of 0.3.",
    "WARNING: clock off its pinned value during oracle default@8: the engine CPUs averaged 3168 MHz, -9.5% against the pinned 3500 MHz; the client CPUs averaged 3214 MHz, -8.2% against the pinned 3500 MHz (limit 1%); this pass is not comparable with passes at the pinned clock."
  ],
  "targets": [
    {
      "name": "sqlitevec",
      "engine": "SQLite 3.53.3 \u002B sqlite-vec 0.1.7-alpha.2.1 (vec0, embedded, in process)",
      "index": "vec0 brute-force scan, no ANN index (exact), float32, cosine distance, default chunk_size=1024; score = 1 - cosine distance; searches run concurrently, one WAL reader connection per searcher (opened as searchers arrive, at most 32), writes run one at a time and may overlap searches",
      "hosting": "embedded",
      "searchSettings": {
        "chunk_size": "1024"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.1226118,
        "rowsPerSecond": 4273.650660050664,
        "countMatchSeconds": 0.0041756,
        "indexSeconds": 0.0042307,
        "indexNote": "nothing to build, every search scans all vectors; checked in 0.00 s: no index, exact scan by design: gvb_gvbbench_eshoponweb is a vec0 virtual table holding 524 vectors and every search compares all of them (EXPLAIN QUERY PLAN: SCAN gvb_gvbbench_eshoponweb VIRTUAL TABLE INDEX 0:3{___}___ | USE TEMP B-TREE FOR ORDER BY)"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 10129,
        "p50Ms": 1.9333,
        "p95Ms": 2.19528,
        "p99Ms": 2.3597199999999994,
        "qps": {
          "1": 506.4432643045848,
          "8": 1138.3659888297357
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 30398,
        "exactP50Ms": 1.9363,
        "exactP95Ms": 2.178,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "1": 2.0665,
          "8": 3.5017
        }
      },
      "ram": {
        "text": "in this process, not measured"
      },
      "disk": {
        "bytes": 10196088,
        "text": "9.72 MiB added by this load (186.34 MiB (engine data folder), 176.61 MiB before)"
      },
      "passOrder": [
        "default@1",
        "default@8",
        "exact"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "totalVectors": 524,
          "detail": "no index, exact scan by design: gvb_gvbbench_eshoponweb is a vec0 virtual table holding 524 vectors and every search compares all of them (EXPLAIN QUERY PLAN: SCAN gvb_gvbbench_eshoponweb VIRTUAL TABLE INDEX 0:3{___}___ | USE TEMP B-TREE FOR ORDER BY)"
        },
        "afterSearch": {
          "ready": true,
          "totalVectors": 524,
          "detail": "no index, exact scan by design: gvb_gvbbench_eshoponweb is a vec0 virtual table holding 524 vectors and every search compares all of them (EXPLAIN QUERY PLAN: SCAN gvb_gvbbench_eshoponweb VIRTUAL TABLE INDEX 0:3{___}___ | USE TEMP B-TREE FOR ORDER BY)"
        }
      },
      "durability": "PRAGMA journal_mode=WAL and PRAGMA synchronous=NORMAL (set by this sink when it opens the file): each commit is appended to the -wal file and handed to the operating system, but the WAL is fsynced only when SQLite checkpoints it (measured with strace: 40 single-record commits caused 3 WAL fsyncs), so a crash of the process loses nothing (measured with kill -9 and a reopen: all 50, 300 and 2,000 rows were there), while an operating-system crash or power cut can lose the newest commits since the last checkpoint and leaves the database consistent",
      "notes": [
        "Index after the load: ready, ? of 524 indexed (no index, exact scan by design: gvb_gvbbench_eshoponweb is a vec0 virtual table holding 524 vectors and every search compares all of them (EXPLAIN QUERY PLAN: SCAN gvb_gvbbench_eshoponweb VIRTUAL TABLE INDEX 0:3{___}___ | USE TEMP B-TREE FOR ORDER BY)). Durability: PRAGMA journal_mode=WAL and PRAGMA synchronous=NORMAL (set by this sink when it opens the file): each commit is appended to the -wal file and handed to the operating system, but the WAL is fsynced only when SQLite checkpoints it (measured with strace: 40 single-record commits caused 3 WAL fsyncs), so a crash of the process loses nothing (measured with kill -9 and a reopen: all 50, 300 and 2,000 rows were there), while an operating-system crash or power cut can lose the newest commits since the last checkpoint and leaves the database consistent.",
        "Outside load when searching began: not known yet (too little sampled history). Load average 4.99 3.36 3.26 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@1 15,028 searches in 30.0 s, default@8 34,255 searches in 30.0 s, exact 15,136 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@1: warm-up 15.0 s and 7,574 searches (0 failed) at 1 searcher; p50 of the last windows 1.944, 1.951, 1.924 ms (windows of at least 2 s and 100 searches); trial of 1,520 searches in 3.0 s at 1 searcher: p50 1.938 ms against the settled p50 1.944 ms, 0% apart (limit 10%); the timed pass p50 1.933 ms, 1% from the settled p50 1.944 ms (limit 10% or 0.1 ms; 0.011 ms); default@8: warm-up 15.0 s and 17,050 searches (0 failed) at 8 searchers; QPS of the last windows 1,141, 1,138, 1,136 (windows of at least 2 s and 100 searches); trial of 3,437 searches in 3.0 s at 8 searchers: 1,144 QPS against the settled 1,138 QPS, 1% apart (limit 10%); the timed pass 1,138 QPS, 0% from the settled 1,138 QPS (limit 10%); exact: warm-up 15.0 s and 7,609 searches (0 failed) at 1 searcher; p50 of the last windows 1.936, 1.934, 1.938 ms (windows of at least 2 s and 100 searches); trial of 1,524 searches in 3.0 s at 1 searcher: p50 1.933 ms against the settled p50 1.936 ms, 0% apart (limit 10%); the timed pass p50 1.936 ms, 0% from the settled p50 1.936 ms (limit 10% or 0.1 ms; 0.000 ms).",
        "Pass default@1 after rehearsal: 2026-10-06T15:50:27.519Z to 2026-10-06T15:50:47.519Z (20.0 s), 10,129 searches, 0 failed, p50 1.933 ms, mean 1.972 ms, p99 2.360 ms, 506.4 QPS (1000/QPS 1.975 ms).",
        "Pass default@8 after default@1: 2026-10-06T15:51:05.588Z to 2026-10-06T15:51:25.591Z (20.0 s), 22,771 searches, 0 failed, p50 6.723 ms, mean 7.016 ms, p99 13.932 ms, 1138.4 QPS (1000/QPS 0.878 ms).",
        "Pass exact after default@8: 2026-10-06T15:51:43.600Z to 2026-10-06T15:52:43.602Z (60.0 s), 30,398 searches, 0 failed, p50 1.936 ms, mean 1.971 ms, p99 2.338 ms, 506.6 QPS (1000/QPS 1.974 ms).",
        "Passes in the order run: default@1, default@8, exact; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 103,133 sent, 0 failed (warmupErrors). Index after the searches: ready, ? of 524 indexed (no index, exact scan by design: gvb_gvbbench_eshoponweb is a vec0 virtual table holding 524 vectors and every search compares all of them (EXPLAIN QUERY PLAN: SCAN gvb_gvbbench_eshoponweb VIRTUAL TABLE INDEX 0:3{___}___ | USE TEMP B-TREE FOR ORDER BY)).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: in this process (embedded), no network: no network address. Open after the passes: localhost:1433 x2 (not this target).",
        "CPU pinning: none: an embedded engine runs inside the benchmark client process, so it shares the client CPUs, CPUs 0-1,4-5.",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@1 3492/3492 MHz, average 3493.4/3492.8, performance, outside load 0.25 before the warm-up, 0.21 during, client CPU 2.067 ms per search; default@8 3492/3492 MHz, average 3492.7/3492.0, performance, outside load 0.24 before the warm-up, 0.23 during, client CPU 3.502 ms per search; exact 3492/3492 MHz, average 3493.6/3492.9, performance, outside load 0.23 before the warm-up, 0.22 during, client CPU 2.069 ms per search."
      ]
    },
    {
      "name": "mariadb",
      "engine": "MariaDB 11.8.9 (InnoDB \u002B VECTOR INDEX)",
      "index": "VECTOR INDEX (HNSW variant) DISTANCE=cosine, M=16 (no ef_construction setting exists), mhnsw_ef_search=100 per statement (the ef 100 most engines here use, so the search effort matches; MariaDB\u0027s own default is 20; recall@10 at ef 100 falls as the set grows (random 1024-dimension vectors, measured 2026-10-04: 0.99 at 524, 0.89 to 0.92 at 2,000; an earlier run gave about 0.09 at 100,000)), mhnsw_max_cache_size 4G; exact mode = IGNORE INDEX full scan",
      "hosting": "compose",
      "searchSettings": {
        "DISTANCE": "cosine",
        "M": "16",
        "mhnsw_ef_search": "100"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.5232483,
        "rowsPerSecond": 1001.4366028518392,
        "countMatchSeconds": 0.0051788,
        "indexSeconds": 0.5302114,
        "indexNote": "VECTOR INDEX is maintained inside every INSERT, nothing to build; confirmed after 2 poll(s) in 0.5 s (poll 1 of 2 said: not ready: a search at mhnsw_ef_search 100 returned only 402 of 524 requested rows through the index (EXPLAIN key \u0027vec_idx\u0027) (information_schema lists vec_idx as INDEX_TYPE VECTOR, EXPLAIN of the default search uses key vec_idx, the server applied mhnsw_ef_search 100 (asked for 100, read back from the server), a search at that effort returned 402 of 524 requested rows through the index; MariaDB has no per-index row count (the graph is a hidden InnoDB table), so indexed vectors are not reported)): information_schema lists vec_idx as INDEX_TYPE VECTOR, EXPLAIN of the default search uses key vec_idx, the server applied mhnsw_ef_search 100 (asked for 100, read back from the server), a search at that effort returned 520 of 524 requested rows through the index; MariaDB has no per-index row count (the graph is a hidden InnoDB table), so indexed vectors are not reported"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 31037,
        "p50Ms": 0.6334,
        "p95Ms": 0.7224199999999997,
        "p99Ms": 0.8774,
        "qps": {
          "1": 1551.839804412485,
          "8": 5917.8401180512155
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 35899,
        "exactP50Ms": 1.6445,
        "exactP95Ms": 1.8130099999999998,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "1": 0.4817,
          "8": 0.3737
        }
      },
      "ram": {
        "bytes": 178991923,
        "text": "170.7 MiB (docker stats)"
      },
      "disk": {
        "bytes": 23077570,
        "text": "22.01 MiB added by this load (182.79 MiB (whole engine data folder), 160.78 MiB before)"
      },
      "passOrder": [
        "default@1",
        "default@8",
        "exact"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "totalVectors": 524,
          "detail": "information_schema lists vec_idx as INDEX_TYPE VECTOR, EXPLAIN of the default search uses key vec_idx, the server applied mhnsw_ef_search 100 (asked for 100, read back from the server), a search at that effort returned 520 of 524 requested rows through the index; MariaDB has no per-index row count (the graph is a hidden InnoDB table), so indexed vectors are not reported"
        },
        "afterSearch": {
          "ready": true,
          "totalVectors": 524,
          "detail": "information_schema lists vec_idx as INDEX_TYPE VECTOR, EXPLAIN of the default search uses key vec_idx, the server applied mhnsw_ef_search 100 (asked for 100, read back from the server), a search at that effort returned 520 of 524 requested rows through the index; MariaDB has no per-index row count (the graph is a hidden InnoDB table), so indexed vectors are not reported"
        }
      },
      "durability": "innodb_flush_log_at_trx_commit=2 (set in mariadb-bench.compose.yaml for the benchmark\u0027s own container gvbbench-mariadb, and in mariadb.compose.yaml for the daily gvb-mariadb, which has the same server settings): the InnoDB redo log is written to the operating system at every commit but fsynced about once a second, so a crash of the mariadbd process loses nothing, while an operating-system crash or power cut can lose the last second of commits; innodb_doublewrite is on (default), the binary log is off, and the vector graph is an InnoDB table under the same log (settings read from the running server with SHOW VARIABLES; the crash behaviour is InnoDB\u0027s documented behaviour for this setting and was not tested on either container)",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started mariadb-bench.compose.yaml with every container created on CPUs 2-3,6-7: gvbbench-mariadb cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [mariadb-bench.compose.yaml e9a2147e6f91]. This run created the engine from these files.",
        "Index after the load: ready, ? of 524 indexed (information_schema lists vec_idx as INDEX_TYPE VECTOR, EXPLAIN of the default search uses key vec_idx, the server applied mhnsw_ef_search 100 (asked for 100, read back from the server), a search at that effort returned 520 of 524 requested rows through the index; MariaDB has no per-index row count (the graph is a hidden InnoDB table), so indexed vectors are not reported). Durability: innodb_flush_log_at_trx_commit=2 (set in mariadb-bench.compose.yaml for the benchmark\u0027s own container gvbbench-mariadb, and in mariadb.compose.yaml for the daily gvb-mariadb, which has the same server settings): the InnoDB redo log is written to the operating system at every commit but fsynced about once a second, so a crash of the mariadbd process loses nothing, while an operating-system crash or power cut can lose the last second of commits; innodb_doublewrite is on (default), the binary log is off, and the vector graph is an InnoDB table under the same log (settings read from the running server with SHOW VARIABLES; the crash behaviour is InnoDB\u0027s documented behaviour for this setting and was not tested on either container).",
        "WARNING: processes outside the benchmark used 0.44 CPUs on average over the last 1 s and 0.44 over the last few seconds when searching began (limit 0.3): the box was busy as the untimed rehearsal started. Each timed pass still waits for a quiet box before its warm-up, and one that ran busy is flagged \u0027busy box\u0027 on its own. The load average 2.27 3.22 3.25 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@1 46,387 searches in 30.0 s, default@8 178,977 searches in 30.0 s, exact 17,994 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@1: warm-up 15.0 s and 23,433 searches (0 failed) at 1 searcher; p50 of the last windows 0.627, 0.623, 0.624 ms (windows of at least 2 s and 100 searches); trial of 4,714 searches in 3.0 s at 1 searcher: p50 0.625 ms against the settled p50 0.624 ms, 0% apart (limit 10%); the timed pass p50 0.633 ms, 1% from the settled p50 0.624 ms (limit 10% or 0.1 ms; 0.009 ms); default@8: warm-up 15.0 s and 88,464 searches (0 failed) at 8 searchers; QPS of the last windows 5,923, 5,838, 5,948 (windows of at least 2 s and 100 searches); trial of 17,668 searches in 3.0 s at 8 searchers: 5,887 QPS against the settled 5,923 QPS, 1% apart (limit 10%); the timed pass 5,918 QPS, 0% from the settled 5,923 QPS (limit 10%); exact: warm-up 15.0 s and 8,975 searches (0 failed) at 1 searcher; p50 of the last windows 1.631, 1.645, 1.656 ms (windows of at least 2 s and 100 searches); trial of 1,801 searches in 3.0 s at 1 searcher: p50 1.650 ms against the settled p50 1.645 ms, 0% apart (limit 10%); the timed pass p50 1.645 ms, 0% from the settled p50 1.645 ms (limit 10% or 0.1 ms; 0.000 ms).",
        "Pass default@1 after rehearsal: 2026-10-06T15:54:39.440Z to 2026-10-06T15:54:59.440Z (20.0 s), 31,037 searches, 0 failed, p50 0.633 ms, mean 0.643 ms, p99 0.877 ms, 1551.8 QPS (1000/QPS 0.644 ms).",
        "Pass default@8 after default@1: 2026-10-06T15:55:17.476Z to 2026-10-06T15:55:37.477Z (20.0 s), 118,362 searches, 0 failed, p50 1.306 ms, mean 1.350 ms, p99 2.476 ms, 5917.8 QPS (1000/QPS 0.169 ms).",
        "Pass exact after default@8: 2026-10-06T15:55:55.523Z to 2026-10-06T15:56:55.523Z (60.0 s), 35,899 searches, 0 failed, p50 1.645 ms, mean 1.669 ms, p99 2.153 ms, 598.3 QPS (1000/QPS 1.671 ms).",
        "Passes in the order run: default@1, default@8, exact; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 388,413 sent, 0 failed (warmupErrors). Index after the searches: ready, ? of 524 indexed (information_schema lists vec_idx as INDEX_TYPE VECTOR, EXPLAIN of the default search uses key vec_idx, the server applied mhnsw_ef_search 100 (asked for 100, read back from the server), a search at that effort returned 520 of 524 requested rows through the index; MariaDB has no per-index row count (the graph is a hidden InnoDB table), so indexed vectors are not reported).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvbbench-mariadb (1ff5bbe73276) on network gvbbench-mariadb_default at container-ip:3306, not through docker-proxy localhost:13306. Open after the passes: container-ip:3306 x8 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvbbench-mariadb: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvbbench-mariadb: 13 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@1 3492/3492 MHz, average 3492.8/3492.1, performance, outside load 0.18 before the warm-up, 0.20 during, client CPU 0.482 ms per search; default@8 3492/3492 MHz, average 3492.0/3492.0, performance, outside load 0.19 before the warm-up, 0.17 during, client CPU 0.374 ms per search; exact 3492/3492 MHz, average 3493.0/3489.7, performance, outside load 0.18 before the warm-up, 0.20 during, client CPU 0.497 ms per search."
      ]
    },
    {
      "name": "duckdb",
      "engine": "DuckDB 1.5.6 \u002B vss b833341 (HNSW, embedded, in process)",
      "index": "HNSW (vss extension) FLOAT[n] metric=cosine m=16 ef_construction=128, ef_search=100 per connection, persistent (hnsw_enable_experimental_persistence=true, checkpoint_threshold=256MB); exact mode = array_cosine_similarity sequential scan; score = 1 - cosine distance; searches run concurrently, one connection per searcher (opened as searchers arrive, at most 32), writes run one at a time and never overlap a search",
      "hosting": "embedded",
      "searchSettings": {
        "checkpoint_threshold": "256MB",
        "ef_construction": "128",
        "ef_search": "100",
        "hnsw_enable_experimental_persistence": "true",
        "m": "16",
        "metric": "cosine"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.363047,
        "rowsPerSecond": 1443.3392921577647,
        "countMatchSeconds": 0.0029415,
        "indexSeconds": 0.0144844,
        "indexNote": "HNSW is maintained inside every transaction, nothing to build; confirmed in 0.01 s: duckdb_indexes() lists gvb_gvbbench_eshoponweb_hnsw, pragma_hnsw_index_info() counts 524 vectors of 524 rows, EXPLAIN of the default search shows HNSW_INDEX_SCAN on it = True"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 5404,
        "p50Ms": 3.66635,
        "p95Ms": 4.073579999999999,
        "p99Ms": 5.217928000000001,
        "qps": {
          "8": 576.0088999547459,
          "1": 270.15952199881895
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 14028,
        "exactP50Ms": 4.1946,
        "exactP95Ms": 4.93715,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "8": 6.9023,
          "1": 5.0167
        }
      },
      "ram": {
        "text": "in this process, not measured"
      },
      "disk": {
        "bytes": 3040286,
        "text": "2.9 MiB added by this load (1.13 GiB (engine data folder), 1.13 GiB before)"
      },
      "passOrder": [
        "default@8",
        "exact",
        "default@1"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "duckdb_indexes() lists gvb_gvbbench_eshoponweb_hnsw, pragma_hnsw_index_info() counts 524 vectors of 524 rows, EXPLAIN of the default search shows HNSW_INDEX_SCAN on it = True"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "duckdb_indexes() lists gvb_gvbbench_eshoponweb_hnsw, pragma_hnsw_index_info() counts 524 vectors of 524 rows, EXPLAIN of the default search shows HNSW_INDEX_SCAN on it = True"
        }
      },
      "durability": "DuckDB\u0027s write-ahead log is fsynced at every commit and no DuckDbSink setting changes that (checkpoint_threshold=256MB only spaces out the checkpoints that write the database file; measured with strace: 42 WAL fsyncs for 40 single-record commits plus 2 setup statements), so an operating-system crash or power cut loses no committed row; measured with kill -9 right after the last commit and a reopen: all 50, 300 and 2,000 rows and an HNSW index that counted the same number were recovered from the log; not tested and documented by DuckDB: the HNSW index is file-backed only through hnsw_enable_experimental_persistence = true, which this sink turns on, and WAL recovery for such custom indexes is not complete, so a crash during a checkpoint or a later commit can damage the index while the rows survive",
      "notes": [
        "Index after the load: ready, 524 of 524 indexed (duckdb_indexes() lists gvb_gvbbench_eshoponweb_hnsw, pragma_hnsw_index_info() counts 524 vectors of 524 rows, EXPLAIN of the default search shows HNSW_INDEX_SCAN on it = True). Durability: DuckDB\u0027s write-ahead log is fsynced at every commit and no DuckDbSink setting changes that (checkpoint_threshold=256MB only spaces out the checkpoints that write the database file; measured with strace: 42 WAL fsyncs for 40 single-record commits plus 2 setup statements), so an operating-system crash or power cut loses no committed row; measured with kill -9 right after the last commit and a reopen: all 50, 300 and 2,000 rows and an HNSW index that counted the same number were recovered from the log; not tested and documented by DuckDB: the HNSW index is file-backed only through hnsw_enable_experimental_persistence = true, which this sink turns on, and WAL recovery for such custom indexes is not complete, so a crash during a checkpoint or a later commit can damage the index while the rows survive.",
        "Outside load when searching began: not known yet (too little sampled history). Load average 2.68 3.43 3.36 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@8 17,319 searches in 30.0 s, exact 7,099 searches in 30.0 s, default@1 8,150 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@8: warm-up 15.0 s and 8,653 searches (0 failed) at 8 searchers; QPS of the last windows 584, 578, 579 (windows of at least 2 s and 100 searches); trial of 1,741 searches in 3.0 s at 8 searchers: 578 QPS against the settled 579 QPS, 0% apart (limit 10%); the timed pass 576 QPS, 1% from the settled 579 QPS (limit 10%); exact: warm-up 15.0 s and 3,532 searches (0 failed) at 1 searcher; p50 of the last windows 4.154, 4.123, 4.117 ms (windows of at least 2 s and 100 searches); trial of 711 searches in 3.0 s at 1 searcher: p50 4.134 ms against the settled p50 4.123 ms, 0% apart (limit 10%); the timed pass p50 4.195 ms, 2% from the settled p50 4.123 ms (limit 10% or 0.1 ms; 0.071 ms); default@1: warm-up 15.0 s and 4,064 searches (0 failed) at 1 searcher; p50 of the last windows 3.654, 3.651, 3.654 ms (windows of at least 2 s and 100 searches); trial of 813 searches in 3.0 s at 1 searcher: p50 3.661 ms against the settled p50 3.654 ms, 0% apart (limit 10%); the timed pass p50 3.666 ms, 0% from the settled p50 3.654 ms (limit 10% or 0.1 ms; 0.013 ms).",
        "Pass default@8 after rehearsal: 2026-10-06T15:58:45.775Z to 2026-10-06T15:59:05.783Z (20.0 s), 11,525 searches, 0 failed, p50 13.169 ms, mean 13.880 ms, p99 28.543 ms, 576.0 QPS (1000/QPS 1.736 ms).",
        "Pass exact after default@8: 2026-10-06T15:59:23.792Z to 2026-10-06T16:00:23.796Z (60.0 s), 14,028 searches, 0 failed, p50 4.195 ms, mean 4.273 ms, p99 5.752 ms, 233.8 QPS (1000/QPS 4.277 ms).",
        "Pass default@1 after exact: 2026-10-06T16:00:41.813Z to 2026-10-06T16:01:01.816Z (20.0 s), 5,404 searches, 0 failed, p50 3.666 ms, mean 3.698 ms, p99 5.218 ms, 270.2 QPS (1000/QPS 3.702 ms).",
        "Passes in the order run: default@8, exact, default@1; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 52,082 sent, 0 failed (warmupErrors). Index after the searches: ready, 524 of 524 indexed (duckdb_indexes() lists gvb_gvbbench_eshoponweb_hnsw, pragma_hnsw_index_info() counts 524 vectors of 524 rows, EXPLAIN of the default search shows HNSW_INDEX_SCAN on it = True).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: in this process (embedded), no network: no network address. No open connection seen (no open TCP connection was found).",
        "CPU pinning: none: an embedded engine runs inside the benchmark client process, so it shares the client CPUs, CPUs 0-1,4-5.",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@8 3492/3492 MHz, average 3492.6/3492.0, performance, outside load 0.21 before the warm-up, 0.23 during, client CPU 6.902 ms per search; exact 3492/3492 MHz, average 3493.2/3492.1, performance, outside load 0.21 before the warm-up, 0.21 during, client CPU 6.251 ms per search; default@1 3492/3492 MHz, average 3493.2/3492.1, performance, outside load 0.21 before the warm-up, 0.19 during, client CPU 5.017 ms per search."
      ]
    },
    {
      "name": "qdrant",
      "engine": "Qdrant 1.17.0 in container gvb-qdrant (image qdrant/qdrant:v1.17.0@sha256:f1c7272cdac52b38c1a0e89313922d940ba50afd90d593a1605dbbc214e66ffb; cpuset 2-3,6-7 from its creation; 3 search thread(s))",
      "index": "exact scan: the builder\u0027s sink sends exact=true on every search, so no HNSW graph is used whether or not Qdrant has built one (see the index state)",
      "hosting": "compose",
      "searchSettings": {
        "exact": "true"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.0960841,
        "rowsPerSecond": 5453.555791228725,
        "countMatchSeconds": 0.009924,
        "indexSeconds": 1.0164272,
        "indexNote": "status Green, 0 of 524 vectors in HNSW segments (2 segments), waited 1.0 s"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 22368,
        "p50Ms": 0.8812,
        "p95Ms": 1.0064649999999997,
        "p99Ms": 1.3974329999999997,
        "qps": {
          "8": 3287.037176836421,
          "1": 1118.3726781554726
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 0,
        "clientCpuMsPerSearch": {
          "8": 0.5235,
          "1": 0.7228
        }
      },
      "ram": {
        "bytes": 37601935,
        "text": "35.86 MiB (docker stats)"
      },
      "disk": {
        "bytes": 205603871,
        "text": "196.08 MiB (collection folder)"
      },
      "passOrder": [
        "default@8",
        "default@1"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 0,
          "totalVectors": 524,
          "detail": "status Green, optimizer ok, indexed_vectors_count 0 of 524 points, 2 segments, indexing_threshold_kb 10,000, full_scan_threshold_kb 10,000; no index used, exact scan by design (the builder\u0027s sink sends exact=true on every search); no graph was built; searches counted since the collection was created: unfiltered_hnsw 0, unfiltered_plain 0, unfiltered_exact 0"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 0,
          "totalVectors": 524,
          "detail": "status Green, optimizer ok, indexed_vectors_count 0 of 524 points, 2 segments, indexing_threshold_kb 10,000, full_scan_threshold_kb 10,000; no index used, exact scan by design (the builder\u0027s sink sends exact=true on every search); no graph was built; searches counted since the collection was created: unfiltered_hnsw 0, unfiltered_plain 597,104, unfiltered_exact 0"
        }
      },
      "durability": "What was measured (strace -f on the Qdrant server process, 30 single-point REST upserts per setting, every sync call timed against the request that caused it): with wait=true each upsert had exactly one msync(MS_SYNC) of the write-ahead-log segment inside the request, before the reply (30 of 30); with wait=false the 30 upserts were acknowledged within 0.26 s with no sync call, and the first WAL msync (one call covering all 30 records) came 2.7 s after the last reply, followed by the segment-file flushes. So the log is flushed to disk under both settings, but with wait=false the flush comes after the acknowledgement: an operating-system crash or power cut in that gap loses acknowledged writes, while a crash of the Qdrant process alone should not, because the bytes are already in the kernel\u0027s page cache (inferred, not tested). Every upsert here is sent with wait=true in batches of 256 over gRPC; the strace used one point per request, so one flush per batch is inferred, not measured. Segment files are flushed every 5 s; log segments 32 MB, 0 created ahead. /qdrant/config/config.yaml in container gvb-qdrant (the image\u0027s own file; its storage keys are listed) sets: storage.collection.quantization = null, storage.collection.replication_factor = 1, storage.collection.vectors.on_disk = null, storage.collection.write_consistency_factor = 1, storage.hnsw_index.ef_construct = 100, storage.hnsw_index.full_scan_threshold_kb = 10000, storage.hnsw_index.m = 16, storage.hnsw_index.max_indexing_threads = 0, storage.hnsw_index.on_disk = false, storage.hnsw_index.payload_m = null, storage.max_collections = null, storage.node_type = Normal, storage.on_disk_payload = true, storage.optimizers.default_segment_number = 0, storage.optimizers.deleted_threshold = 0.2, storage.optimizers.flush_interval_sec = 5, storage.optimizers.indexing_threshold_kb = 10000, storage.optimizers.max_optimization_threads = null, storage.optimizers.max_segment_size_kb = null, storage.optimizers.vacuum_min_vector_number = 1000, storage.performance.max_search_threads = 0, storage.performance.optimizer_cpu_budget = 0, storage.performance.update_rate_limit = null, storage.shard_transfer_method = null, storage.snapshots_config.snapshots_storage = local, storage.snapshots_path = ./snapshots, storage.storage_path = ./storage, storage.temp_path = null, storage.update_concurrency = null, storage.wal.wal_capacity_mb = 32, storage.wal.wal_segments_ahead = 0; no QDRANT__ environment overrides in the server process. Not tested by cutting power; whether the disk\u0027s own write cache reaches the media was not checked.",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started qdrant.compose.yaml with every container created on CPUs 2-3,6-7: gvb-qdrant cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [qdrant.compose.yaml eb57e55e93c1]. This run created the engine from these files.",
        "Index after the load: ready, 0 of 524 indexed (status Green, optimizer ok, indexed_vectors_count 0 of 524 points, 2 segments, indexing_threshold_kb 10,000, full_scan_threshold_kb 10,000; no index used, exact scan by design (the builder\u0027s sink sends exact=true on every search); no graph was built; searches counted since the collection was created: unfiltered_hnsw 0, unfiltered_plain 0, unfiltered_exact 0). Durability: What was measured (strace -f on the Qdrant server process, 30 single-point REST upserts per setting, every sync call timed against the request that caused it): with wait=true each upsert had exactly one msync(MS_SYNC) of the write-ahead-log segment inside the request, before the reply (30 of 30); with wait=false the 30 upserts were acknowledged within 0.26 s with no sync call, and the first WAL msync (one call covering all 30 records) came 2.7 s after the last reply, followed by the segment-file flushes. So the log is flushed to disk under both settings, but with wait=false the flush comes after the acknowledgement: an operating-system crash or power cut in that gap loses acknowledged writes, while a crash of the Qdrant process alone should not, because the bytes are already in the kernel\u0027s page cache (inferred, not tested). Every upsert here is sent with wait=true in batches of 256 over gRPC; the strace used one point per request, so one flush per batch is inferred, not measured. Segment files are flushed every 5 s; log segments 32 MB, 0 created ahead. /qdrant/config/config.yaml in container gvb-qdrant (the image\u0027s own file; its storage keys are listed) sets: storage.collection.quantization = null, storage.collection.replication_factor = 1, storage.collection.vectors.on_disk = null, storage.collection.write_consistency_factor = 1, storage.hnsw_index.ef_construct = 100, storage.hnsw_index.full_scan_threshold_kb = 10000, storage.hnsw_index.m = 16, storage.hnsw_index.max_indexing_threads = 0, storage.hnsw_index.on_disk = false, storage.hnsw_index.payload_m = null, storage.max_collections = null, storage.node_type = Normal, storage.on_disk_payload = true, storage.optimizers.default_segment_number = 0, storage.optimizers.deleted_threshold = 0.2, storage.optimizers.flush_interval_sec = 5, storage.optimizers.indexing_threshold_kb = 10000, storage.optimizers.max_optimization_threads = null, storage.optimizers.max_segment_size_kb = null, storage.optimizers.vacuum_min_vector_number = 1000, storage.performance.max_search_threads = 0, storage.performance.optimizer_cpu_budget = 0, storage.performance.update_rate_limit = null, storage.shard_transfer_method = null, storage.snapshots_config.snapshots_storage = local, storage.snapshots_path = ./snapshots, storage.storage_path = ./storage, storage.temp_path = null, storage.update_concurrency = null, storage.wal.wal_capacity_mb = 32, storage.wal.wal_segments_ahead = 0; no QDRANT__ environment overrides in the server process. Not tested by cutting power; whether the disk\u0027s own write cache reaches the media was not checked..",
        "Outside load when searching began: 0.19 CPUs on average over the last 1 s and 0.19 over the last few seconds (limit 0.3), a quiet box. The load average 2.74 4.01 3.71 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@8 97,950 searches in 30.0 s, default@1 33,458 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@8: warm-up 15.0 s and 49,125 searches (0 failed) at 8 searchers; QPS of the last windows 3,284, 3,283, 3,272 (windows of at least 2 s and 100 searches); trial of 9,790 searches in 3.0 s at 8 searchers: 3,261 QPS against the settled 3,283 QPS, 1% apart (limit 10%); the timed pass 3,287 QPS, 0% from the settled 3,283 QPS (limit 10%); default@1: warm-up 15.0 s and 16,745 searches (0 failed) at 1 searcher; p50 of the last windows 0.881, 0.879, 0.878 ms (windows of at least 2 s and 100 searches); trial of 3,368 searches in 3.0 s at 1 searcher: p50 0.878 ms against the settled p50 0.879 ms, 0% apart (limit 10%); the timed pass p50 0.881 ms, 0% from the settled p50 0.879 ms (limit 10% or 0.1 ms; 0.002 ms).",
        "Pass default@8 after rehearsal: 2026-10-06T16:02:27.715Z to 2026-10-06T16:02:47.718Z (20.0 s), 65,748 searches, 0 failed, p50 2.353 ms, mean 2.431 ms, p99 4.198 ms, 3287.0 QPS (1000/QPS 0.304 ms).",
        "Pass default@1 after default@8: 2026-10-06T16:03:05.742Z to 2026-10-06T16:03:25.743Z (20.0 s), 22,368 searches, 0 failed, p50 0.881 ms, mean 0.892 ms, p99 1.397 ms, 1118.4 QPS (1000/QPS 0.894 ms).",
        "Passes in the order run: default@8, default@1; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 210,436 sent, 0 failed (warmupErrors). Index after the searches: ready, 0 of 524 indexed (status Green, optimizer ok, indexed_vectors_count 0 of 524 points, 2 segments, indexing_threshold_kb 10,000, full_scan_threshold_kb 10,000; no index used, exact scan by design (the builder\u0027s sink sends exact=true on every search); no graph was built; searches counted since the collection was created: unfiltered_hnsw 0, unfiltered_plain 597,104, unfiltered_exact 0).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-qdrant (e08d8b70ccba) on network gvb-qdrant_default at container-ip:6334, not through docker-proxy localhost:16334; container gvb-qdrant (e08d8b70ccba) on network gvb-qdrant_default at container-ip:6333, not through docker-proxy localhost:16333. Open after the passes: container-ip:6334 x1 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-qdrant: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-qdrant: 24 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@8 3492/3492 MHz, average 3492.0/3492.0, performance, outside load 0.21 before the warm-up, 0.20 during, client CPU 0.524 ms per search; default@1 3492/3492 MHz, average 3492.0/3492.0, performance, outside load 0.20 before the warm-up, 0.17 during, client CPU 0.723 ms per search."
      ]
    },
    {
      "name": "elasticsearch",
      "engine": "Elasticsearch 9.5.3 (dense_vector)",
      "index": "HNSW float32, no quantization, m=16, ef_construction=128, cosine; search k=top, num_candidates=100; 1 shard, 0 replicas; force-merged to one segment after the load (at 1,024 dimensions a segment under 1,043 vectors gets no graph)",
      "hosting": "compose",
      "searchSettings": {
        "ef_construction": "128",
        "k": "top",
        "m": "16",
        "num_candidates": "100"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 1.3413824,
        "rowsPerSecond": 390.64177374028463,
        "countMatchSeconds": 0.1173134,
        "indexSeconds": 0.3707256,
        "indexNote": "FAILED after 0.4 s: NO HNSW GRAPH, searches scan all 524 vectors: 1 segment(s), 524 vectors, total_vex_size_bytes 0 (_stats dense_vector), index_options hnsw m=16 ef_construction=128; Lucene builds no graph for a segment this small (measured at 1,024 dimensions: 1,042 vectors none, 1,043 vectors a graph), so default search equals exact search"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 14962,
        "p50Ms": 1.3102,
        "p95Ms": 1.5125899999999997,
        "p99Ms": 1.7531119999999996,
        "qps": {
          "8": 2707.9679162373122,
          "1": 748.0569381023581
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 46677,
        "exactP50Ms": 1.2607,
        "exactP95Ms": 1.4646,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "8": 0.4909,
          "1": 0.4584
        }
      },
      "ram": {
        "bytes": 2694018236,
        "text": "2.51 GiB (docker stats)"
      },
      "disk": {
        "bytes": 2499235,
        "text": "2.38 MiB added by this load (2.55 MiB (whole engine data folder), 172.25 KiB before)"
      },
      "passOrder": [
        "default@8",
        "exact",
        "default@1"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": false,
          "indexedVectors": 0,
          "totalVectors": 524,
          "detail": "NO HNSW GRAPH, searches scan all 524 vectors: 1 segment(s), 524 vectors, total_vex_size_bytes 0 (_stats dense_vector), index_options hnsw m=16 ef_construction=128; Lucene builds no graph for a segment this small (measured at 1,024 dimensions: 1,042 vectors none, 1,043 vectors a graph), so default search equals exact search"
        },
        "afterSearch": {
          "ready": false,
          "indexedVectors": 0,
          "totalVectors": 524,
          "detail": "NO HNSW GRAPH, searches scan all 524 vectors: 1 segment(s), 524 vectors, total_vex_size_bytes 0 (_stats dense_vector), index_options hnsw m=16 ef_construction=128; Lucene builds no graph for a segment this small (measured at 1,024 dimensions: 1,042 vectors none, 1,043 vectors a graph), so default search equals exact search"
        }
      },
      "durability": "Every acknowledged bulk request is fsynced to the translog before the answer: index.translog.durability=request, the Elasticsearch default, which neither this sink nor elasticsearch.compose.yaml overrides (ElasticsearchReadinessTests reads it back from the live index). By that setting a process crash or power loss loses no acknowledged write; this is read from the setting, not shown by pulling power. One node and no replicas, so a lost disk loses the data.",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started elasticsearch.compose.yaml with every container created on CPUs 2-3,6-7: gvb-elasticsearch cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [elasticsearch.compose.yaml 582b9a97b8de]. This run created the engine from these files.",
        "WARNING: the index step after the load FAILED (NO HNSW GRAPH, searches scan all 524 vectors: 1 segment(s), 524 vectors, total_vex_size_bytes 0 (_stats dense_vector), index_options hnsw m=16 ef_construction=128; Lucene builds no graph for a segment this small (measured at 1,024 dimensions: 1,042 vectors none, 1,043 vectors a graph), so default search equals exact search). Searched anyway; see the index state.",
        "WARNING: index not ready after the load: NOT ready, 0 of 524 indexed (NO HNSW GRAPH, searches scan all 524 vectors: 1 segment(s), 524 vectors, total_vex_size_bytes 0 (_stats dense_vector), index_options hnsw m=16 ef_construction=128; Lucene builds no graph for a segment this small (measured at 1,024 dimensions: 1,042 vectors none, 1,043 vectors a graph), so default search equals exact search). Measured anyway, so its search numbers may come from a scan or a half-built index. Durability: Every acknowledged bulk request is fsynced to the translog before the answer: index.translog.durability=request, the Elasticsearch default, which neither this sink nor elasticsearch.compose.yaml overrides (ElasticsearchReadinessTests reads it back from the live index). By that setting a process crash or power loss loses no acknowledged write; this is read from the setting, not shown by pulling power. One node and no replicas, so a lost disk loses the data..",
        "Outside load when searching began: 0.19 CPUs on average over the last 2 s and 0.19 over the last few seconds (limit 0.3), a quiet box. The load average 3.01 3.89 3.74 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@8 37,772 searches in 30.0 s, exact 21,967 searches in 30.0 s, default@1 22,526 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@8: warm-up 15.0 s and 39,619 searches (0 failed) at 8 searchers; QPS of the last windows 2,704, 2,702, 2,702 (windows of at least 2 s and 100 searches); trial of 8,139 searches in 3.0 s at 8 searchers: 2,712 QPS against the settled 2,702 QPS, 0% apart (limit 10%); the timed pass 2,708 QPS, 0% from the settled 2,702 QPS (limit 10%); exact: warm-up 15.0 s and 11,661 searches (0 failed) at 1 searcher; p50 of the last windows 1.254, 1.259, 1.258 ms (windows of at least 2 s and 100 searches); trial of 2,361 searches in 3.0 s at 1 searcher: p50 1.253 ms against the settled p50 1.258 ms, 0% apart (limit 10%); the timed pass p50 1.261 ms, 0% from the settled p50 1.258 ms (limit 10% or 0.1 ms; 0.003 ms); default@1: warm-up 15.0 s and 11,239 searches (0 failed) at 1 searcher; p50 of the last windows 1.306, 1.309, 1.311 ms (windows of at least 2 s and 100 searches); trial of 2,262 searches in 3.0 s at 1 searcher: p50 1.305 ms against the settled p50 1.309 ms, 0% apart (limit 10%); the timed pass p50 1.310 ms, 0% from the settled p50 1.309 ms (limit 10% or 0.1 ms; 0.002 ms).",
        "Pass default@8 after rehearsal: 2026-10-06T16:05:48.819Z to 2026-10-06T16:06:08.821Z (20.0 s), 54,163 searches, 0 failed, p50 2.830 ms, mean 2.952 ms, p99 5.340 ms, 2708.0 QPS (1000/QPS 0.369 ms).",
        "Pass exact after default@8: 2026-10-06T16:06:26.840Z to 2026-10-06T16:07:26.841Z (60.0 s), 46,677 searches, 0 failed, p50 1.261 ms, mean 1.283 ms, p99 1.679 ms, 777.9 QPS (1000/QPS 1.285 ms).",
        "Pass default@1 after exact: 2026-10-06T16:07:44.888Z to 2026-10-06T16:08:04.889Z (20.0 s), 14,962 searches, 0 failed, p50 1.310 ms, mean 1.335 ms, p99 1.753 ms, 748.1 QPS (1000/QPS 1.337 ms).",
        "Passes in the order run: default@8, exact, default@1; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 157,546 sent, 0 failed (warmupErrors). Index after the searches: NOT ready, 0 of 524 indexed (NO HNSW GRAPH, searches scan all 524 vectors: 1 segment(s), 524 vectors, total_vex_size_bytes 0 (_stats dense_vector), index_options hnsw m=16 ef_construction=128; Lucene builds no graph for a segment this small (measured at 1,024 dimensions: 1,042 vectors none, 1,043 vectors a graph), so default search equals exact search).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-elasticsearch (2fe4965f42f7) on network engines_default at container-ip:9200, not through docker-proxy localhost:9200. Open after the passes: container-ip:9200 x1 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-elasticsearch: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-elasticsearch: 75 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@8 3492/3492 MHz, average 3492.0/3492.2, performance, outside load 0.21 before the warm-up, 0.25 during, client CPU 0.491 ms per search; exact 3492/3492 MHz, average 3491.9/3492.3, performance, outside load 0.23 before the warm-up, 0.23 during, client CPU 0.461 ms per search; default@1 3492/3492 MHz, average 3491.9/3491.9, performance, outside load 0.23 before the warm-up, 0.22 during, client CPU 0.458 ms per search."
      ]
    },
    {
      "name": "vespa",
      "engine": "Vespa 8.754.14 (tensor attribute \u002B HNSW)",
      "index": "HNSW float32 tensor, prenormalized-angular (cosine), max-links-per-node=16, neighbors-to-explore-at-insert=128; search targetHits=top, ef=100 via exploreAdditionalHits; exact mode = approximate:false; vectors held in memory",
      "hosting": "compose",
      "searchSettings": {
        "ef": "100",
        "max-links-per-node": "16",
        "neighbors-to-explore-at-insert": "128",
        "targetHits": "top"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 1.6761091,
        "rowsPerSecond": 312.6288139596641,
        "countMatchSeconds": 0.0129228,
        "indexSeconds": 0.0488626,
        "indexNote": "nothing to build, Vespa inserts into the HNSW graph while writing; ready after 0.0 s: HNSW graph returns 524 of 524 vectors: totalCount 524, coverage full True, nearestNeighbor approximate:true with targetHits above the count: has_index True, algorithm \u0022index top k\u0022, top_k_hits 524; the graph is updated while each write is applied, so there is no later build step"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 9919,
        "p50Ms": 1.9484,
        "p95Ms": 2.4224,
        "p99Ms": 2.7263959999999994,
        "qps": {
          "1": 495.944053630797,
          "8": 1564.7377511477457
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 29861,
        "exactP50Ms": 1.9348,
        "exactP95Ms": 2.43,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "1": 0.9373,
          "8": 0.8754
        }
      },
      "ram": {
        "bytes": 2978559819,
        "text": "2.77 GiB (docker stats)"
      },
      "disk": {
        "bytes": 12368088,
        "text": "11.8 MiB added by this load (2.13 GiB (whole engine data folder), 2.11 GiB before)"
      },
      "passOrder": [
        "default@1",
        "default@8",
        "exact"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "HNSW graph returns 524 of 524 vectors: totalCount 524, coverage full True, nearestNeighbor approximate:true with targetHits above the count: has_index True, algorithm \u0022index top k\u0022, top_k_hits 524; the graph is updated while each write is applied, so there is no later build step"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "HNSW graph returns 524 of 524 vectors: totalCount 524, coverage full True, nearestNeighbor approximate:true with targetHits above the count: has_index True, algorithm \u0022index top k\u0022, top_k_hits 524; the graph is updated while each write is applied, so there is no later build step"
        }
      },
      "durability": "Vespa\u0027s transaction log server fsyncs after each commit: searchlib.translogserver usefsync=true, and an operation is searchable when it is acknowledged: proton documentdb visibilitydelay=0. Neither is overridden by this sink or by the services.xml it deploys; VespaReadinessTests reads both from the live config server. After a crash the node replays the transaction log over its last flushed data and rebuilds the in-memory HNSW graph. By those settings an acknowledged write survives a crash; this is read from the settings, not shown by pulling power. One node and min-redundancy 1, so a lost disk loses the data.",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started vespa.compose.yaml with every container created on CPUs 2-3,6-7: gvb-vespa cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [vespa.compose.yaml 3cfb702cc29f]. This run created the engine from these files.",
        "Index after the load: ready, 524 of 524 indexed (HNSW graph returns 524 of 524 vectors: totalCount 524, coverage full True, nearestNeighbor approximate:true with targetHits above the count: has_index True, algorithm \u0022index top k\u0022, top_k_hits 524; the graph is updated while each write is applied, so there is no later build step). Durability: Vespa\u0027s transaction log server fsyncs after each commit: searchlib.translogserver usefsync=true, and an operation is searchable when it is acknowledged: proton documentdb visibilitydelay=0. Neither is overridden by this sink or by the services.xml it deploys; VespaReadinessTests reads both from the live config server. After a crash the node replays the transaction log over its last flushed data and rebuilds the in-memory HNSW graph. By those settings an acknowledged write survives a crash; this is read from the settings, not shown by pulling power. One node and min-redundancy 1, so a lost disk loses the data..",
        "Outside load when searching began: 0.23 CPUs on average over the last 40 s and 0.24 over the last few seconds (limit 0.3), a quiet box. The load average 4.41 3.82 3.77 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@1 8,285 searches in 30.0 s, default@8 35,286 searches in 30.0 s, exact 14,806 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@1: warm-up 15.0 s and 6,952 searches (0 failed) at 1 searcher; p50 of the last windows 2.009, 2.017, 1.946 ms (windows of at least 2 s and 100 searches); trial of 1,488 searches in 3.0 s at 1 searcher: p50 1.941 ms against the settled p50 2.009 ms, 4% apart (limit 10%); the timed pass p50 1.948 ms, 3% from the settled p50 2.009 ms (limit 10% or 0.1 ms; 0.061 ms); default@8: warm-up 15.0 s and 23,237 searches (0 failed) at 8 searchers; QPS of the last windows 1,558, 1,564, 1,552 (windows of at least 2 s and 100 searches); trial of 4,697 searches in 3.0 s at 8 searchers: 1,564 QPS against the settled 1,558 QPS, 0% apart (limit 10%); the timed pass 1,565 QPS, 0% from the settled 1,558 QPS (limit 10%); exact: warm-up 15.0 s and 7,394 searches (0 failed) at 1 searcher; p50 of the last windows 1.945, 1.942, 1.949 ms (windows of at least 2 s and 100 searches); trial of 1,450 searches in 3.0 s at 1 searcher: p50 1.944 ms against the settled p50 1.945 ms, 0% apart (limit 10%); the timed pass p50 1.935 ms, 1% from the settled p50 1.945 ms (limit 10% or 0.1 ms; 0.010 ms).",
        "Pass default@1 after rehearsal: 2026-10-06T16:10:52.850Z to 2026-10-06T16:11:12.851Z (20.0 s), 9,919 searches, 0 failed, p50 1.948 ms, mean 2.014 ms, p99 2.726 ms, 495.9 QPS (1000/QPS 2.016 ms).",
        "Pass default@8 after default@1: 2026-10-06T16:11:30.866Z to 2026-10-06T16:11:50.869Z (20.0 s), 31,299 searches, 0 failed, p50 4.987 ms, mean 5.110 ms, p99 8.310 ms, 1564.7 QPS (1000/QPS 0.639 ms).",
        "Pass exact after default@8: 2026-10-06T16:12:08.881Z to 2026-10-06T16:13:08.883Z (60.0 s), 29,861 searches, 0 failed, p50 1.935 ms, mean 2.006 ms, p99 2.801 ms, 497.7 QPS (1000/QPS 2.009 ms).",
        "Passes in the order run: default@1, default@8, exact; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 103,595 sent, 0 failed (warmupErrors). Index after the searches: ready, 524 of 524 indexed (HNSW graph returns 524 of 524 vectors: totalCount 524, coverage full True, nearestNeighbor approximate:true with targetHits above the count: has_index True, algorithm \u0022index top k\u0022, top_k_hits 524; the graph is updated while each write is applied, so there is no later build step).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-vespa (862c328d26b8) on network engines_default at container-ip:8080, not through docker-proxy localhost:8090; container gvb-vespa (862c328d26b8) on network engines_default at container-ip:19071, not through docker-proxy localhost:19071. Open after the passes: container-ip:8080 x1 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-vespa: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-vespa: 214 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@1 3492/3492 MHz, average 3492.0/3492.2, performance, outside load 0.23 before the warm-up, 0.20 during, client CPU 0.937 ms per search; default@8 3492/3492 MHz, average 3492.0/3492.1, performance, outside load 0.19 before the warm-up, 0.26 during, client CPU 0.875 ms per search; exact 3492/3492 MHz, average 3492.0/3492.2, performance, outside load 0.24 before the warm-up, 0.18 during, client CPU 0.952 ms per search."
      ]
    },
    {
      "name": "qdrant-hnsw",
      "engine": "Qdrant 1.17.0 in container gvb-qdrant (image qdrant/qdrant:v1.17.0@sha256:f1c7272cdac52b38c1a0e89313922d940ba50afd90d593a1605dbbc214e66ffb; cpuset 2-3,6-7 from its creation; 3 search thread(s))",
      "index": "HNSW m=16 ef_construct=100, hnsw_ef=server default, cosine; indexing_threshold_kb 1 and full_scan_threshold_kb 10 (server defaults are 10,000 each) so a small collection builds and walks its graph",
      "hosting": "compose",
      "searchSettings": {
        "ef_construct": "100",
        "hnsw_ef": "server",
        "m": "16"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.0741438,
        "rowsPerSecond": 7067.347505792798,
        "countMatchSeconds": 0.0028526,
        "indexSeconds": 2.0084898,
        "indexNote": "indexing_threshold_kb 1, full_scan_threshold_kb 10; status Green, 524 of 524 vectors in HNSW segments (2 segments), waited 2.0 s"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 21546,
        "p50Ms": 0.9163,
        "p95Ms": 1.0356,
        "p99Ms": 1.433904999999999,
        "qps": {
          "1": 1077.2869648277256,
          "8": 3448.268759302981
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 66375,
        "exactP50Ms": 0.8912,
        "exactP95Ms": 1.0151299999999994,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "1": 0.7302,
          "8": 0.5116
        }
      },
      "ram": {
        "bytes": 39678115,
        "text": "37.84 MiB (docker stats)"
      },
      "disk": {
        "bytes": 172067450,
        "text": "164.1 MiB (collection folder)"
      },
      "passOrder": [
        "exact",
        "default@1",
        "default@8"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "status Green, optimizer ok, indexed_vectors_count 524 of 524 points, 2 segments, indexing_threshold_kb 1, full_scan_threshold_kb 10; every one of 1 non-empty segments has a complete HNSW graph above its full-scan threshold; no search has run yet, so the walk is proven by the segment settings only; searches counted since the collection was created: unfiltered_hnsw 0, unfiltered_plain 0, unfiltered_exact 0"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "status Green, optimizer ok, indexed_vectors_count 524 of 524 points, 2 segments, indexing_threshold_kb 1, full_scan_threshold_kb 10; every one of 1 non-empty segments has a complete HNSW graph above its full-scan threshold; 307,551 segment searches walked the graph and none scanned; searches counted since the collection was created: unfiltered_hnsw 307,551, unfiltered_plain 0, unfiltered_exact 119,913"
        }
      },
      "durability": "What was measured (strace -f on the Qdrant server process, 30 single-point REST upserts per setting, every sync call timed against the request that caused it): with wait=true each upsert had exactly one msync(MS_SYNC) of the write-ahead-log segment inside the request, before the reply (30 of 30); with wait=false the 30 upserts were acknowledged within 0.26 s with no sync call, and the first WAL msync (one call covering all 30 records) came 2.7 s after the last reply, followed by the segment-file flushes. So the log is flushed to disk under both settings, but with wait=false the flush comes after the acknowledgement: an operating-system crash or power cut in that gap loses acknowledged writes, while a crash of the Qdrant process alone should not, because the bytes are already in the kernel\u0027s page cache (inferred, not tested). Every upsert here is sent with wait=true in batches of 256 over gRPC; the strace used one point per request, so one flush per batch is inferred, not measured. Segment files are flushed every 5 s; log segments 32 MB, 0 created ahead. /qdrant/config/config.yaml in container gvb-qdrant (the image\u0027s own file; its storage keys are listed) sets: storage.collection.quantization = null, storage.collection.replication_factor = 1, storage.collection.vectors.on_disk = null, storage.collection.write_consistency_factor = 1, storage.hnsw_index.ef_construct = 100, storage.hnsw_index.full_scan_threshold_kb = 10000, storage.hnsw_index.m = 16, storage.hnsw_index.max_indexing_threads = 0, storage.hnsw_index.on_disk = false, storage.hnsw_index.payload_m = null, storage.max_collections = null, storage.node_type = Normal, storage.on_disk_payload = true, storage.optimizers.default_segment_number = 0, storage.optimizers.deleted_threshold = 0.2, storage.optimizers.flush_interval_sec = 5, storage.optimizers.indexing_threshold_kb = 10000, storage.optimizers.max_optimization_threads = null, storage.optimizers.max_segment_size_kb = null, storage.optimizers.vacuum_min_vector_number = 1000, storage.performance.max_search_threads = 0, storage.performance.optimizer_cpu_budget = 0, storage.performance.update_rate_limit = null, storage.shard_transfer_method = null, storage.snapshots_config.snapshots_storage = local, storage.snapshots_path = ./snapshots, storage.storage_path = ./storage, storage.temp_path = null, storage.update_concurrency = null, storage.wal.wal_capacity_mb = 32, storage.wal.wal_segments_ahead = 0; no QDRANT__ environment overrides in the server process. Not tested by cutting power; whether the disk\u0027s own write cache reaches the media was not checked.",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started qdrant.compose.yaml with every container created on CPUs 2-3,6-7: gvb-qdrant cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [qdrant.compose.yaml eb57e55e93c1]. This run created the engine from these files.",
        "Index after the load: ready, 524 of 524 indexed (status Green, optimizer ok, indexed_vectors_count 524 of 524 points, 2 segments, indexing_threshold_kb 1, full_scan_threshold_kb 10; every one of 1 non-empty segments has a complete HNSW graph above its full-scan threshold; no search has run yet, so the walk is proven by the segment settings only; searches counted since the collection was created: unfiltered_hnsw 0, unfiltered_plain 0, unfiltered_exact 0). Durability: What was measured (strace -f on the Qdrant server process, 30 single-point REST upserts per setting, every sync call timed against the request that caused it): with wait=true each upsert had exactly one msync(MS_SYNC) of the write-ahead-log segment inside the request, before the reply (30 of 30); with wait=false the 30 upserts were acknowledged within 0.26 s with no sync call, and the first WAL msync (one call covering all 30 records) came 2.7 s after the last reply, followed by the segment-file flushes. So the log is flushed to disk under both settings, but with wait=false the flush comes after the acknowledgement: an operating-system crash or power cut in that gap loses acknowledged writes, while a crash of the Qdrant process alone should not, because the bytes are already in the kernel\u0027s page cache (inferred, not tested). Every upsert here is sent with wait=true in batches of 256 over gRPC; the strace used one point per request, so one flush per batch is inferred, not measured. Segment files are flushed every 5 s; log segments 32 MB, 0 created ahead. /qdrant/config/config.yaml in container gvb-qdrant (the image\u0027s own file; its storage keys are listed) sets: storage.collection.quantization = null, storage.collection.replication_factor = 1, storage.collection.vectors.on_disk = null, storage.collection.write_consistency_factor = 1, storage.hnsw_index.ef_construct = 100, storage.hnsw_index.full_scan_threshold_kb = 10000, storage.hnsw_index.m = 16, storage.hnsw_index.max_indexing_threads = 0, storage.hnsw_index.on_disk = false, storage.hnsw_index.payload_m = null, storage.max_collections = null, storage.node_type = Normal, storage.on_disk_payload = true, storage.optimizers.default_segment_number = 0, storage.optimizers.deleted_threshold = 0.2, storage.optimizers.flush_interval_sec = 5, storage.optimizers.indexing_threshold_kb = 10000, storage.optimizers.max_optimization_threads = null, storage.optimizers.max_segment_size_kb = null, storage.optimizers.vacuum_min_vector_number = 1000, storage.performance.max_search_threads = 0, storage.performance.optimizer_cpu_budget = 0, storage.performance.update_rate_limit = null, storage.shard_transfer_method = null, storage.snapshots_config.snapshots_storage = local, storage.snapshots_path = ./snapshots, storage.storage_path = ./storage, storage.temp_path = null, storage.update_concurrency = null, storage.wal.wal_capacity_mb = 32, storage.wal.wal_segments_ahead = 0; no QDRANT__ environment overrides in the server process. Not tested by cutting power; whether the disk\u0027s own write cache reaches the media was not checked..",
        "Outside load when searching began: 0.18 CPUs on average over the last 2 s and 0.18 over the last few seconds (limit 0.3), a quiet box. The load average 3.71 4.43 4.09 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: exact 33,652 searches in 30.0 s, default@1 32,892 searches in 30.0 s, default@8 103,233 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. exact: warm-up 15.0 s and 16,565 searches (0 failed) at 1 searcher; p50 of the last windows 0.893, 0.896, 0.892 ms (windows of at least 2 s and 100 searches); trial of 3,321 searches in 3.0 s at 1 searcher: p50 0.890 ms against the settled p50 0.893 ms, 0% apart (limit 10%); the timed pass p50 0.891 ms, 0% from the settled p50 0.893 ms (limit 10% or 0.1 ms; 0.002 ms); default@1: warm-up 15.0 s and 16,121 searches (0 failed) at 1 searcher; p50 of the last windows 0.921, 0.921, 0.918 ms (windows of at least 2 s and 100 searches); trial of 3,226 searches in 3.0 s at 1 searcher: p50 0.919 ms against the settled p50 0.921 ms, 0% apart (limit 10%); the timed pass p50 0.916 ms, 0% from the settled p50 0.921 ms (limit 10% or 0.1 ms; 0.004 ms); default@8: warm-up 15.0 s and 51,220 searches (0 failed) at 8 searchers; QPS of the last windows 3,384, 3,467, 3,411 (windows of at least 2 s and 100 searches); trial of 10,341 searches in 3.0 s at 8 searchers: 3,445 QPS against the settled 3,411 QPS, 1% apart (limit 10%); the timed pass 3,448 QPS, 1% from the settled 3,411 QPS (limit 10%).",
        "Pass exact after rehearsal: 2026-10-06T16:15:14.953Z to 2026-10-06T16:16:14.953Z (60.0 s), 66,375 searches, 0 failed, p50 0.891 ms, mean 0.902 ms, p99 1.396 ms, 1106.2 QPS (1000/QPS 0.904 ms).",
        "Pass default@1 after exact: 2026-10-06T16:16:33.022Z to 2026-10-06T16:16:53.022Z (20.0 s), 21,546 searches, 0 failed, p50 0.916 ms, mean 0.927 ms, p99 1.434 ms, 1077.3 QPS (1000/QPS 0.928 ms).",
        "Pass default@8 after default@1: 2026-10-06T16:17:11.046Z to 2026-10-06T16:17:31.048Z (20.0 s), 68,972 searches, 0 failed, p50 2.234 ms, mean 2.318 ms, p99 3.999 ms, 3448.3 QPS (1000/QPS 0.290 ms).",
        "Passes in the order run: exact, default@1, default@8; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 270,571 sent, 0 failed (warmupErrors). Index after the searches: ready, 524 of 524 indexed (status Green, optimizer ok, indexed_vectors_count 524 of 524 points, 2 segments, indexing_threshold_kb 1, full_scan_threshold_kb 10; every one of 1 non-empty segments has a complete HNSW graph above its full-scan threshold; 307,551 segment searches walked the graph and none scanned; searches counted since the collection was created: unfiltered_hnsw 307,551, unfiltered_plain 0, unfiltered_exact 119,913).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-qdrant (fa2ccc8859f4) on network gvb-qdrant_default at container-ip:6334, not through docker-proxy localhost:16334; container gvb-qdrant (fa2ccc8859f4) on network gvb-qdrant_default at container-ip:6333, not through docker-proxy localhost:16333. Open after the passes: container-ip:6334 x1 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-qdrant: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-qdrant: 24 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): exact 3492/3492 MHz, average 3492.0/3492.0, performance, outside load 0.19 before the warm-up, 0.18 during, client CPU 0.726 ms per search; default@1 3492/3492 MHz, average 3492.0/3491.8, performance, outside load 0.18 before the warm-up, 0.16 during, client CPU 0.730 ms per search; default@8 3492/3492 MHz, average 3492.0/3492.0, performance, outside load 0.17 before the warm-up, 0.21 during, client CPU 0.512 ms per search."
      ]
    },
    {
      "name": "sql-diskann",
      "engine": "Microsoft SQL Server 2025 (RTM-CU9) (KB5122048) - 17.0.5005.3 (X64), Enterprise Developer Edition (64-bit), in container gvb-mssql (image mcr.microsoft.com/mssql/server:2025-CU9-ubuntu-24.04@sha256:2b5b581621126574f3d1f75e78d3eebe8d05aedb59ad0cfdf9aa42cb0634d726; cpuset 2-3,6-7 from its creation; SQL Server counts 4 CPU(s) and runs 4 visible scheduler(s), affinity AUTO)",
      "index": "DiskANN (preview) via VECTOR_SEARCH, cosine, build {\u0022StartId\u0022:\u0022306\u0022, \u0022L\u0022:\u002248\u0022, \u0022M\u0022:\u00228\u0022, \u0022R\u0022:\u002248\u0022}; the exact mode scans the same table",
      "hosting": "compose",
      "searchSettings": {
        "L": "48",
        "M": "8",
        "R": "48",
        "StartId": "306"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.9341081,
        "rowsPerSecond": 560.9629121083524,
        "countMatchSeconds": 0.0038891,
        "indexSeconds": 3.9762131,
        "indexNote": "copied 524 rows to dbo.gvb_gvbbench_eshoponweb_ann (INT key) and built DiskANN, parameters {\u0022StartId\u0022:\u0022306\u0022, \u0022L\u0022:\u002248\u0022, \u0022M\u0022:\u00228\u0022, \u0022R\u0022:\u002248\u0022}, graph covers 524 rows"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 5359,
        "p50Ms": 3.6845,
        "p95Ms": 4.169009999999999,
        "p99Ms": 4.758732,
        "qps": {
          "8": 645.3751213907675,
          "1": 267.9145736559255
        },
        "recall": 0.9650000000000002,
        "ndcg": 0.5256995680516536,
        "errors": 0,
        "exactQueries": 15442,
        "exactP50Ms": 3.78955,
        "exactP95Ms": 4.297274999999999,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "8": 1.0818,
          "1": 1.0236
        }
      },
      "ram": {
        "bytes": 554906419,
        "text": "529.2 MiB (docker stats)"
      },
      "disk": {
        "bytes": 4734976,
        "text": "4.52 MiB (table and its indexes, reserved pages)"
      },
      "passOrder": [
        "default@8",
        "default@1",
        "exact"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "DiskANN index built and used. sys.vector_indexes: vix_gvb_gvbbench_eshoponweb on dbo.gvb_gvbbench_eshoponweb_ann in GvbBenchDiskAnn, DiskANN, COSINE, enabled, build {\u0022StartId\u0022:\u0022306\u0022, \u0022L\u0022:\u002248\u0022, \u0022M\u0022:\u00228\u0022, \u0022R\u0022:\u002248\u0022}; graph table rows 524 of 524 table rows. Real VECTOR_SEARCH plan: Vector Index Seek on index vix_gvb_gvbbench_eshoponweb of GvbBenchDiskAnn.gvb_gvbbench_eshoponweb_ann (IndexKind DiskANN), 10 rows returned, 10 hits. The exact mode searches the SAME table (dbo.gvb_gvbbench_eshoponweb_ann, database GvbBenchDiskAnn) and its real plan is: Clustered Index Scan of GvbBenchDiskAnn.gvb_gvbbench_eshoponweb_ann through PK_gvb_gvbbench_eshoponweb_ann (IndexKind Clustered), 524 rows read, no vector index operator, 10 hits"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "DiskANN index built and used. sys.vector_indexes: vix_gvb_gvbbench_eshoponweb on dbo.gvb_gvbbench_eshoponweb_ann in GvbBenchDiskAnn, DiskANN, COSINE, enabled, build {\u0022StartId\u0022:\u0022306\u0022, \u0022L\u0022:\u002248\u0022, \u0022M\u0022:\u00228\u0022, \u0022R\u0022:\u002248\u0022}; graph table rows 524 of 524 table rows. Real VECTOR_SEARCH plan: Vector Index Seek on index vix_gvb_gvbbench_eshoponweb of GvbBenchDiskAnn.gvb_gvbbench_eshoponweb_ann (IndexKind DiskANN), 10 rows returned, 10 hits. The exact mode searches the SAME table (dbo.gvb_gvbbench_eshoponweb_ann, database GvbBenchDiskAnn) and its real plan is: Clustered Index Scan of GvbBenchDiskAnn.gvb_gvbbench_eshoponweb_ann through PK_gvb_gvbbench_eshoponweb_ann (IndexKind Clustered), 524 rows read, no vector index operator, 10 hits"
        }
      },
      "durability": "A commit returns after its transaction-log records are written to disk (SQL Server write-ahead logging; delayed durability is DISABLED); a database created here copies the model database: recovery model FULL, page_verify CHECKSUM; no global trace flags are enabled; mssql.conf of container gvb-mssql (/var/opt/mssql/mssql.conf, on the host at ~/gvb-data/engines/mssql/mssql.conf) does not exist, so it sets: nothing; it has no [control] or [traceflag] entry, so SQL Server\u0027s own Linux defaults for flushing writes apply (Microsoft\u0027s Linux performance guide names trace flag 3982 as that default; read from the guide, not tested here). Not tested by cutting power; whether the disk\u0027s own write cache reaches the media was not checked. Container settings from its environment (names only): MSSQL_AGENT_ENABLED, MSSQL_MEMORY_LIMIT_MB, MSSQL_PID, MSSQL_RPC_PORT, MSSQL_SA_PASSWORD.",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started mssql.compose.yaml with every container created on CPUs 2-3,6-7: gvb-mssql cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [mssql.compose.yaml b7359663cc15]. This run created the engine from these files.",
        "Index after the load: ready, 524 of 524 indexed (DiskANN index built and used. sys.vector_indexes: vix_gvb_gvbbench_eshoponweb on dbo.gvb_gvbbench_eshoponweb_ann in GvbBenchDiskAnn, DiskANN, COSINE, enabled, build {\u0022StartId\u0022:\u0022306\u0022, \u0022L\u0022:\u002248\u0022, \u0022M\u0022:\u00228\u0022, \u0022R\u0022:\u002248\u0022}; graph table rows 524 of 524 table rows. Real VECTOR_SEARCH plan: Vector Index Seek on index vix_gvb_gvbbench_eshoponweb of GvbBenchDiskAnn.gvb_gvbbench_eshoponweb_ann (IndexKind DiskANN), 10 rows returned, 10 hits. The exact mode searches the SAME table (dbo.gvb_gvbbench_eshoponweb_ann, database GvbBenchDiskAnn) and its real plan is: Clustered Index Scan of GvbBenchDiskAnn.gvb_gvbbench_eshoponweb_ann through PK_gvb_gvbbench_eshoponweb_ann (IndexKind Clustered), 524 rows read, no vector index operator, 10 hits). Durability: A commit returns after its transaction-log records are written to disk (SQL Server write-ahead logging; delayed durability is DISABLED); a database created here copies the model database: recovery model FULL, page_verify CHECKSUM; no global trace flags are enabled; mssql.conf of container gvb-mssql (/var/opt/mssql/mssql.conf, on the host at ~/gvb-data/engines/mssql/mssql.conf) does not exist, so it sets: nothing; it has no [control] or [traceflag] entry, so SQL Server\u0027s own Linux defaults for flushing writes apply (Microsoft\u0027s Linux performance guide names trace flag 3982 as that default; read from the guide, not tested here). Not tested by cutting power; whether the disk\u0027s own write cache reaches the media was not checked. Container settings from its environment (names only): MSSQL_AGENT_ENABLED, MSSQL_MEMORY_LIMIT_MB, MSSQL_PID, MSSQL_RPC_PORT, MSSQL_SA_PASSWORD..",
        "Outside load when searching began: 0.25 CPUs on average over the last 6 s and 0.24 over the last few seconds (limit 0.3), a quiet box. The load average 4.11 3.92 3.94 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@8 20,367 searches in 30.0 s, default@1 8,188 searches in 30.0 s, exact 7,808 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@8: warm-up 15.0 s and 10,275 searches (0 failed) at 8 searchers; QPS of the last windows 668, 695, 672 (windows of at least 2 s and 100 searches); trial of 2,085 searches in 3.0 s at 8 searchers: 692 QPS against the settled 672 QPS, 3% apart (limit 10%); the timed pass 645 QPS, 4% from the settled 672 QPS (limit 10%); default@1: warm-up 15.0 s and 4,025 searches (0 failed) at 1 searcher; p50 of the last windows 3.685, 3.670, 3.676 ms (windows of at least 2 s and 100 searches); trial of 804 searches in 3.0 s at 1 searcher: p50 3.673 ms against the settled p50 3.676 ms, 0% apart (limit 10%); the timed pass p50 3.685 ms, 0% from the settled p50 3.676 ms (limit 10% or 0.1 ms; 0.008 ms); exact: warm-up 15.0 s and 3,884 searches (0 failed) at 1 searcher; p50 of the last windows 3.747, 3.784, 3.775 ms (windows of at least 2 s and 100 searches); trial of 772 searches in 3.0 s at 1 searcher: p50 3.802 ms against the settled p50 3.775 ms, 1% apart (limit 10%); the timed pass p50 3.790 ms, 0% from the settled p50 3.775 ms (limit 10% or 0.1 ms; 0.015 ms).",
        "Pass default@8 after rehearsal: 2026-10-06T16:19:35.209Z to 2026-10-06T16:19:55.219Z (20.0 s), 12,914 searches, 0 failed, p50 11.433 ms, mean 12.390 ms, p99 23.458 ms, 645.4 QPS (1000/QPS 1.549 ms).",
        "Pass default@1 after default@8: 2026-10-06T16:20:13.227Z to 2026-10-06T16:20:33.230Z (20.0 s), 5,359 searches, 0 failed, p50 3.685 ms, mean 3.730 ms, p99 4.759 ms, 267.9 QPS (1000/QPS 3.733 ms).",
        "Pass exact after default@1: 2026-10-06T16:20:51.239Z to 2026-10-06T16:21:51.242Z (60.0 s), 15,442 searches, 0 failed, p50 3.790 ms, mean 3.882 ms, p99 5.252 ms, 257.4 QPS (1000/QPS 3.886 ms).",
        "Passes in the order run: default@8, default@1, exact; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 58,208 sent, 0 failed (warmupErrors). Index after the searches: ready, 524 of 524 indexed (DiskANN index built and used. sys.vector_indexes: vix_gvb_gvbbench_eshoponweb on dbo.gvb_gvbbench_eshoponweb_ann in GvbBenchDiskAnn, DiskANN, COSINE, enabled, build {\u0022StartId\u0022:\u0022306\u0022, \u0022L\u0022:\u002248\u0022, \u0022M\u0022:\u00228\u0022, \u0022R\u0022:\u002248\u0022}; graph table rows 524 of 524 table rows. Real VECTOR_SEARCH plan: Vector Index Seek on index vix_gvb_gvbbench_eshoponweb of GvbBenchDiskAnn.gvb_gvbbench_eshoponweb_ann (IndexKind DiskANN), 10 rows returned, 10 hits. The exact mode searches the SAME table (dbo.gvb_gvbbench_eshoponweb_ann, database GvbBenchDiskAnn) and its real plan is: Clustered Index Scan of GvbBenchDiskAnn.gvb_gvbbench_eshoponweb_ann through PK_gvb_gvbbench_eshoponweb_ann (IndexKind Clustered), 524 rows read, no vector index operator, 10 hits).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-mssql (8b2960fc1293) on network gvb-mssql_default at container-ip:1433, not through docker-proxy localhost:14330. Open after the passes: container-ip:1433 x9 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-mssql: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-mssql: 84 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@8 3492/3492 MHz, average 3492.1/3492.5, performance, outside load 0.21 before the warm-up, 0.20 during, client CPU 1.082 ms per search; default@1 3492/3492 MHz, average 3492.4/3492.4, performance, outside load 0.19 before the warm-up, 0.23 during, client CPU 1.024 ms per search; exact 3492/3492 MHz, average 3492.4/3492.3, performance, outside load 0.22 before the warm-up, 0.21 during, client CPU 1.155 ms per search."
      ]
    },
    {
      "name": "pgvector",
      "engine": "PostgreSQL 17 \u002B pgvector 0.8.7",
      "index": "HNSW vector_cosine_ops m=16 ef_construction=128, hnsw.ef_search=100 per query, float32 vector(n), cosine; exact mode = same query with index scans off (sequential scan)",
      "hosting": "compose",
      "searchSettings": {
        "ef_construction": "128",
        "hnsw.ef_search": "100",
        "m": "16"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.8572834,
        "rowsPerSecond": 611.233111477488,
        "countMatchSeconds": 0.0115763,
        "indexSeconds": 0.0261199,
        "indexNote": "HNSW is maintained inside every insert, nothing to build; confirmed in 0.0 s: pg_indexes lists gvb_gvbbench_eshoponweb_hnsw (USING hnsw (embedding vector_cosine_ops) WITH (m=\u002716\u0027, ef_construction=\u0027128\u0027)), pg_index valid and ready = True, EXPLAIN of the default search uses Index Scan on it = True, that search returned 10 of 10 rows; PostgreSQL keeps no entry count for an HNSW index, so indexed vectors are not reported"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 22348,
        "p50Ms": 0.8768,
        "p95Ms": 1.0460649999999998,
        "p99Ms": 1.2903589999999996,
        "qps": {
          "8": 3909.9337726094227,
          "1": 1117.3561605310415
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 26089,
        "exactP50Ms": 2.2516,
        "exactP95Ms": 2.55892,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "8": 0.5333,
          "1": 0.6688
        }
      },
      "ram": {
        "bytes": 105172172,
        "text": "100.3 MiB (docker stats)"
      },
      "disk": {
        "bytes": 7929856,
        "text": "7.56 MiB added by this load (374.57 MiB (whole engine data folder), 367 MiB before)"
      },
      "passOrder": [
        "exact",
        "default@8",
        "default@1"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "totalVectors": 524,
          "detail": "pg_indexes lists gvb_gvbbench_eshoponweb_hnsw (USING hnsw (embedding vector_cosine_ops) WITH (m=\u002716\u0027, ef_construction=\u0027128\u0027)), pg_index valid and ready = True, EXPLAIN of the default search uses Index Scan on it = True, that search returned 10 of 10 rows; PostgreSQL keeps no entry count for an HNSW index, so indexed vectors are not reported"
        },
        "afterSearch": {
          "ready": true,
          "totalVectors": 524,
          "detail": "pg_indexes lists gvb_gvbbench_eshoponweb_hnsw (USING hnsw (embedding vector_cosine_ops) WITH (m=\u002716\u0027, ef_construction=\u0027128\u0027)), pg_index valid and ready = True, EXPLAIN of the default search uses Index Scan on it = True, that search returned 10 of 10 rows; PostgreSQL keeps no entry count for an HNSW index, so indexed vectors are not reported"
        }
      },
      "durability": "fsync on, synchronous_commit on, full_page_writes on, wal_sync_method fdatasync (PostgreSQL defaults; pgvector.compose.yaml sets only shared_buffers, maintenance_work_mem and max_wal_size): every commit is flushed to the write-ahead log before it returns and HNSW index changes are WAL-logged, so a crash loses no committed row (max_wal_size 4GB only spaces out checkpoints; measured with docker kill, which is SIGKILL, right after a 2,000-vector load and a restart: all 2,000 rows were there and the HNSW index was valid and used by the default search)",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started pgvector.compose.yaml with every container created on CPUs 2-3,6-7: gvb-pgvector cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [pgvector.compose.yaml 27549563c622]. This run created the engine from these files.",
        "Index after the load: ready, ? of 524 indexed (pg_indexes lists gvb_gvbbench_eshoponweb_hnsw (USING hnsw (embedding vector_cosine_ops) WITH (m=\u002716\u0027, ef_construction=\u0027128\u0027)), pg_index valid and ready = True, EXPLAIN of the default search uses Index Scan on it = True, that search returned 10 of 10 rows; PostgreSQL keeps no entry count for an HNSW index, so indexed vectors are not reported). Durability: fsync on, synchronous_commit on, full_page_writes on, wal_sync_method fdatasync (PostgreSQL defaults; pgvector.compose.yaml sets only shared_buffers, maintenance_work_mem and max_wal_size): every commit is flushed to the write-ahead log before it returns and HNSW index changes are WAL-logged, so a crash loses no committed row (max_wal_size 4GB only spaces out checkpoints; measured with docker kill, which is SIGKILL, right after a 2,000-vector load and a restart: all 2,000 rows were there and the HNSW index was valid and used by the default search).",
        "Outside load when searching began: 0.24 CPUs on average over the last 1 s and 0.24 over the last few seconds (limit 0.3), a quiet box. The load average 1.90 2.98 3.57 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: exact 13,565 searches in 30.0 s, default@8 122,921 searches in 30.0 s, default@1 34,098 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. exact: warm-up 15.0 s and 6,495 searches (0 failed) at 1 searcher; p50 of the last windows 2.274, 2.258, 2.248 ms (windows of at least 2 s and 100 searches); trial of 1,299 searches in 3.0 s at 1 searcher: p50 2.268 ms against the settled p50 2.258 ms, 0% apart (limit 10%); the timed pass p50 2.252 ms, 0% from the settled p50 2.258 ms (limit 10% or 0.1 ms; 0.007 ms); default@8: warm-up 15.0 s and 58,764 searches (0 failed) at 8 searchers; QPS of the last windows 3,858, 3,948, 3,884 (windows of at least 2 s and 100 searches); trial of 11,868 searches in 3.0 s at 8 searchers: 3,954 QPS against the settled 3,884 QPS, 2% apart (limit 10%); the timed pass 3,910 QPS, 1% from the settled 3,884 QPS (limit 10%); default@1: warm-up 15.0 s and 16,655 searches (0 failed) at 1 searcher; p50 of the last windows 0.879, 0.886, 0.875 ms (windows of at least 2 s and 100 searches); trial of 3,327 searches in 3.0 s at 1 searcher: p50 0.887 ms against the settled p50 0.879 ms, 1% apart (limit 10%); the timed pass p50 0.877 ms, 0% from the settled p50 0.879 ms (limit 10% or 0.1 ms; 0.002 ms).",
        "Pass exact after rehearsal: 2026-10-06T16:23:48.721Z to 2026-10-06T16:24:48.723Z (60.0 s), 26,089 searches, 0 failed, p50 2.252 ms, mean 2.298 ms, p99 3.334 ms, 434.8 QPS (1000/QPS 2.300 ms).",
        "Pass default@8 after exact: 2026-10-06T16:25:06.749Z to 2026-10-06T16:25:26.751Z (20.0 s), 78,204 searches, 0 failed, p50 1.991 ms, mean 2.044 ms, p99 3.698 ms, 3909.9 QPS (1000/QPS 0.256 ms).",
        "Pass default@1 after default@8: 2026-10-06T16:25:44.779Z to 2026-10-06T16:26:04.780Z (20.0 s), 22,348 searches, 0 failed, p50 0.877 ms, mean 0.893 ms, p99 1.290 ms, 1117.4 QPS (1000/QPS 0.895 ms).",
        "Passes in the order run: exact, default@8, default@1; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 268,992 sent, 0 failed (warmupErrors). Index after the searches: ready, ? of 524 indexed (pg_indexes lists gvb_gvbbench_eshoponweb_hnsw (USING hnsw (embedding vector_cosine_ops) WITH (m=\u002716\u0027, ef_construction=\u0027128\u0027)), pg_index valid and ready = True, EXPLAIN of the default search uses Index Scan on it = True, that search returned 10 of 10 rows; PostgreSQL keeps no entry count for an HNSW index, so indexed vectors are not reported).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-pgvector (a107b4cc5b1b) on network gvb-pgvector_default at container-ip:5432, not through docker-proxy localhost:5432. Open after the passes: container-ip:5432 x8 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-pgvector: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-pgvector: 6 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): exact 3492/3492 MHz, average 3492.2/3492.4, performance, outside load 0.21 before the warm-up, 0.20 during, client CPU 0.691 ms per search; default@8 3492/3492 MHz, average 3492.0/3492.0, performance, outside load 0.20 before the warm-up, 0.24 during, client CPU 0.533 ms per search; default@1 3492/3492 MHz, average 3492.4/3492.1, performance, outside load 0.23 before the warm-up, 0.21 during, client CPU 0.669 ms per search."
      ]
    },
    {
      "name": "weaviate",
      "engine": "Weaviate 1.39.8 (single node, REST \u002B GraphQL)",
      "index": "HNSW maxConnections(M)=16 efConstruction=128, ef=-1 (dynamic: limit x 8 clamped 100..500), cosine, no quantization; approximate only (no exact mode)",
      "hosting": "compose",
      "searchSettings": {
        "ef": "-1",
        "efConstruction": "128"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.5519868,
        "rowsPerSecond": 949.2980629246931,
        "countMatchSeconds": 0.0043674,
        "indexSeconds": 0.0182041,
        "indexNote": "HNSW is updated inside every batch, nothing to build; confirmed in 0.0 s: shard gtd945wzSIS6 vectorIndexingStatus READY, vectorQueueLength 0, node status objectCount 0 (refreshed only when the memtable is flushed, so it lags a minute and does not decide readiness), schema shard gtd945wzSIS6 status READY, Aggregate count 524, Aggregate nearVector with objectLimit 524 reached 524, Weaviate reports no count of vectors in the HNSW index, so indexed vectors are not reported"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 3309,
        "p50Ms": 5.3722,
        "p95Ms": 10.9018,
        "p99Ms": 11.971364,
        "qps": {
          "1": 165.4404549129538,
          "8": 305.39734408358635
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 0,
        "clientCpuMsPerSearch": {
          "1": 0.5952,
          "8": 0.6226
        }
      },
      "ram": {
        "bytes": 99866378,
        "text": "95.24 MiB (docker stats)"
      },
      "disk": {
        "bytes": 3102153,
        "text": "2.96 MiB added by this load (5.02 MiB (whole engine data folder), 2.06 MiB before)"
      },
      "passOrder": [
        "default@1",
        "default@8"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "totalVectors": 524,
          "detail": "shard gtd945wzSIS6 vectorIndexingStatus READY, vectorQueueLength 0, node status objectCount 0 (refreshed only when the memtable is flushed, so it lags a minute and does not decide readiness), schema shard gtd945wzSIS6 status READY, Aggregate count 524, Aggregate nearVector with objectLimit 524 reached 524, Weaviate reports no count of vectors in the HNSW index, so indexed vectors are not reported"
        },
        "afterSearch": {
          "ready": true,
          "totalVectors": 524,
          "detail": "shard gtd945wzSIS6 vectorIndexingStatus READY, vectorQueueLength 0, node status objectCount 524 (refreshed only when the memtable is flushed, so it lags a minute and does not decide readiness), schema shard gtd945wzSIS6 status READY, Aggregate count 524, Aggregate nearVector with objectLimit 524 reached 524, Weaviate reports no count of vectors in the HNSW index, so indexed vectors are not reported"
        }
      },
      "durability": "Weaviate 1.39.8 defaults, weaviate.compose.yaml sets no persistence variable: every object is appended to the LSM write-ahead log with a plain write and no fsync, and the log is fsynced only when its memtable is flushed, 60 seconds after the last write (PERSISTENCE_MEMTABLES_FLUSH_IDLE_AFTER_SECONDS default; measured with strace: 2,080 writes into the objects log during a 2,000-object load, the first fsync 60 s after the last write), so a power cut can lose the last minute of writes; the HNSW commit log is buffered inside the process (77 writes for 2,000 vectors), so a killed process loses the newest vectors from the vector index while their objects survive (measured with docker kill, which is SIGKILL, one second after the load and a restart, two runs each: 0 to 1 of 50, 264 to 267 of 300 and 1,992 to 1,997 of 2,000 stored objects were still reachable through a vector search)",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started weaviate.compose.yaml with every container created on CPUs 2-3,6-7: gvb-weaviate cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [weaviate.compose.yaml d96b7399662a]. This run created the engine from these files.",
        "Index after the load: ready, ? of 524 indexed (shard gtd945wzSIS6 vectorIndexingStatus READY, vectorQueueLength 0, node status objectCount 0 (refreshed only when the memtable is flushed, so it lags a minute and does not decide readiness), schema shard gtd945wzSIS6 status READY, Aggregate count 524, Aggregate nearVector with objectLimit 524 reached 524, Weaviate reports no count of vectors in the HNSW index, so indexed vectors are not reported). Durability: Weaviate 1.39.8 defaults, weaviate.compose.yaml sets no persistence variable: every object is appended to the LSM write-ahead log with a plain write and no fsync, and the log is fsynced only when its memtable is flushed, 60 seconds after the last write (PERSISTENCE_MEMTABLES_FLUSH_IDLE_AFTER_SECONDS default; measured with strace: 2,080 writes into the objects log during a 2,000-object load, the first fsync 60 s after the last write), so a power cut can lose the last minute of writes; the HNSW commit log is buffered inside the process (77 writes for 2,000 vectors), so a killed process loses the newest vectors from the vector index while their objects survive (measured with docker kill, which is SIGKILL, one second after the load and a restart, two runs each: 0 to 1 of 50, 264 to 267 of 300 and 1,992 to 1,997 of 2,000 stored objects were still reachable through a vector search).",
        "WARNING: processes outside the benchmark used 0.36 CPUs on average over the last 1 s and 0.36 over the last few seconds when searching began (limit 0.3): the box was busy as the untimed rehearsal started. Each timed pass still waits for a quiet box before its warm-up, and one that ran busy is flagged \u0027busy box\u0027 on its own. The load average 2.58 3.23 3.55 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@1 5,021 searches in 30.0 s, default@8 9,276 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@1: warm-up 15.0 s and 2,492 searches (0 failed) at 1 searcher; p50 of the last windows 5.371, 5.374, 5.397 ms (windows of at least 2 s and 100 searches); trial of 497 searches in 3.0 s at 1 searcher: p50 5.353 ms against the settled p50 5.374 ms, 0% apart (limit 10%); the timed pass p50 5.372 ms, 0% from the settled p50 5.374 ms (limit 10% or 0.1 ms; 0.001 ms); default@8: warm-up 15.0 s and 4,582 searches (0 failed) at 8 searchers; QPS of the last windows 307, 304, 306 (windows of at least 2 s and 100 searches); trial of 924 searches in 3.0 s at 8 searchers: 306 QPS against the settled 306 QPS, 0% apart (limit 10%); the timed pass 305 QPS, 0% from the settled 306 QPS (limit 10%).",
        "Pass default@1 after rehearsal: 2026-10-06T16:27:46.811Z to 2026-10-06T16:28:06.812Z (20.0 s), 3,309 searches, 0 failed, p50 5.372 ms, mean 6.041 ms, p99 11.971 ms, 165.4 QPS (1000/QPS 6.044 ms).",
        "Pass default@8 after default@1: 2026-10-06T16:28:24.852Z to 2026-10-06T16:28:44.862Z (20.0 s), 6,111 searches, 0 failed, p50 25.086 ms, mean 26.186 ms, p99 54.411 ms, 305.4 QPS (1000/QPS 3.274 ms).",
        "Passes in the order run: default@1, default@8; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 22,792 sent, 0 failed (warmupErrors). Index after the searches: ready, ? of 524 indexed (shard gtd945wzSIS6 vectorIndexingStatus READY, vectorQueueLength 0, node status objectCount 524 (refreshed only when the memtable is flushed, so it lags a minute and does not decide readiness), schema shard gtd945wzSIS6 status READY, Aggregate count 524, Aggregate nearVector with objectLimit 524 reached 524, Weaviate reports no count of vectors in the HNSW index, so indexed vectors are not reported).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-weaviate (f60c5682a178) on network gvb-weaviate_default at container-ip:8080, not through docker-proxy localhost:8085. Open after the passes: container-ip:8080 x8 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-weaviate: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-weaviate: 9 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@1 3492/3492 MHz, average 3492.1/3492.6, performance, outside load 0.22 before the warm-up, 0.21 during, client CPU 0.595 ms per search; default@8 3492/3492 MHz, average 3492.0/3492.3, performance, outside load 0.21 before the warm-up, 0.21 during, client CPU 0.623 ms per search."
      ]
    },
    {
      "name": "clickhouse",
      "engine": "ClickHouse 26.3.39.7 (MergeTree \u002B vector_similarity index; image-default system logging, each log table with a TTL of 1 hour)",
      "index": "vector_similarity HNSW cosineDistance, quantization bf16, M=16 ef_construction=128, hnsw_candidate_list_size_for_search=256, rescoring off; exact mode = full scan with skip indexes off",
      "hosting": "compose",
      "searchSettings": {
        "M": "16",
        "ef_construction": "128",
        "hnsw_candidate_list_size_for_search": "256"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.2240708,
        "rowsPerSecond": 2338.54656653165,
        "countMatchSeconds": 0.0025865,
        "indexSeconds": 6.1804637,
        "indexNote": "1 data part(s), index files on 1 of them, 524 of 524 live rows in indexed parts; 0 merge(s) running, 0 unfinished mutation(s); plan of the default search uses the Skip index vec_idx on 1 of 1 parts; wait took 6.2 s"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 3862,
        "p50Ms": 4.9567,
        "p95Ms": 7.200155,
        "p99Ms": 7.933881999999998,
        "qps": {
          "8": 402.40999809529933,
          "1": 193.08674169888124
        },
        "recall": 0.99,
        "ndcg": 0.514829573372199,
        "errors": 0,
        "exactQueries": 8752,
        "exactP50Ms": 6.6774,
        "exactP95Ms": 7.6775899999999995,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "8": 1.0624,
          "1": 0.8856
        }
      },
      "ram": {
        "bytes": 1068498944,
        "text": "1019 MiB (docker stats)"
      },
      "disk": {
        "bytes": 0,
        "text": "0 B added by this load (6.79 GiB (whole engine data folder), 6.79 GiB before)"
      },
      "passOrder": [
        "default@8",
        "exact",
        "default@1"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "1 data part(s), index files on 1 of them, 524 of 524 live rows in indexed parts; 0 merge(s) running, 0 unfinished mutation(s); plan of the default search uses the Skip index vec_idx on 1 of 1 parts"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "1 data part(s), index files on 1 of them, 524 of 524 live rows in indexed parts; 0 merge(s) running, 0 unfinished mutation(s); plan of the default search uses the Skip index vec_idx on 1 of 1 parts"
        }
      },
      "durability": "Acknowledged inserts are not fsynced. The sink creates plain MergeTree tables, so the MergeTree settings are the defaults: fsync_after_insert 0 and fsync_part_directory 0 (system.merge_tree_settings), and async_insert 1 with wait_for_async_insert 1 is the server default for the gvb login (system.settings), so an insert is acknowledged once its part is written, not once it is on disk. Measured 2026-10-04 with strace over 30 acknowledged single-row inserts (30 Ok rows in system.asynchronous_insert_log): zero fsync, fdatasync or sync_file_range calls, while the control, a table created with fsync_after_insert=1, made 61 fdatasync calls for 5 inserts. A host power loss or kernel crash can lose acknowledged rows that are still in the OS page cache; a ClickHouse process crash alone should not (inferred, not tested).",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started clickhouse.compose.yaml with every container created on CPUs 2-3,6-7: gvb-clickhouse cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [clickhouse-config/gvb-system-logs.xml 02cc53f0f171; clickhouse.compose.yaml 560e0ae487c6]. This run created the engine from these files.",
        "Index after the load: ready, 524 of 524 indexed (1 data part(s), index files on 1 of them, 524 of 524 live rows in indexed parts; 0 merge(s) running, 0 unfinished mutation(s); plan of the default search uses the Skip index vec_idx on 1 of 1 parts). Durability: Acknowledged inserts are not fsynced. The sink creates plain MergeTree tables, so the MergeTree settings are the defaults: fsync_after_insert 0 and fsync_part_directory 0 (system.merge_tree_settings), and async_insert 1 with wait_for_async_insert 1 is the server default for the gvb login (system.settings), so an insert is acknowledged once its part is written, not once it is on disk. Measured 2026-10-04 with strace over 30 acknowledged single-row inserts (30 Ok rows in system.asynchronous_insert_log): zero fsync, fdatasync or sync_file_range calls, while the control, a table created with fsync_after_insert=1, made 61 fdatasync calls for 5 inserts. A host power loss or kernel crash can lose acknowledged rows that are still in the OS page cache; a ClickHouse process crash alone should not (inferred, not tested)..",
        "Outside load when searching began: 0.25 CPUs on average over the last 7 s and 0.25 over the last few seconds (limit 0.3), a quiet box. The load average 3.00 3.17 3.48 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@8 12,897 searches in 30.0 s, exact 4,357 searches in 30.0 s, default@1 5,760 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@8: warm-up 15.0 s and 6,442 searches (0 failed) at 8 searchers; QPS of the last windows 466, 470, 394 (windows of at least 2 s and 100 searches) (its last 3 windows differed by more than 5%); trial of 1,129 searches in 3.0 s at 8 searchers: 375 QPS against the settled 466 QPS, 24% apart (limit 10%), so the warm-up was EXTENDED once (30.0 s and 12,206 searches (0 failed) at 8 searchers; QPS of its older 6 windows 420 (windows 378, 379, 447, 470, 456, 387) and of its newer 6 402 (windows 399, 463, 470, 393, 344, 344), 4.3% apart (limit 5%) (windows of at least 2 s and 100 searches)); second trial of 1,261 searches in 3.0 s at 8 searchers: 418 QPS against the settled 402 QPS, 4% apart (limit 10%), inside its newer windows\u0027 range 344 to 470 QPS; settled after the extension; the timed pass 402 QPS, 0% from the settled 402 QPS (limit 10%); exact: warm-up 15.0 s and 2,088 searches (0 failed) at 1 searcher; p50 of the last windows 7.199, 6.671, 6.603 ms (windows of at least 2 s and 100 searches) (its last 3 windows differed by more than 5%); trial of 441 searches in 3.0 s at 1 searcher: p50 6.665 ms against the settled p50 6.671 ms, 0% apart (limit 10%), so the warm-up was EXTENDED once (30.0 s and 3,866 searches (0 failed) at 1 searcher; p50 of its older 6 windows 6.876 ms (windows 6.911, 6.885, 6.851, 6.846, 6.867, 6.848) and of its newer 6 6.976 ms (windows 7.041, 6.942, 6.890, 7.208, 7.080, 6.888), 1.5% apart (limit 5% or 0.1 ms; 0.100 ms) (windows of at least 2 s and 100 searches)); second trial of 430 searches in 3.0 s at 1 searcher: p50 6.728 ms against the settled p50 6.976 ms, 4% apart (limit 10%), outside its newer windows\u0027 range 6.888 to 7.208 ms; settled after the extension; the timed pass p50 6.677 ms, 4% from the settled p50 6.976 ms (limit 10% or 0.1 ms; 0.298 ms); default@1: warm-up 15.0 s and 2,932 searches (0 failed) at 1 searcher; p50 of the last windows 4.901, 4.947, 4.925 ms (windows of at least 2 s and 100 searches); trial of 574 searches in 3.0 s at 1 searcher: p50 5.004 ms against the settled p50 4.925 ms, 2% apart (limit 10%); the timed pass p50 4.957 ms, 1% from the settled p50 4.925 ms (limit 10% or 0.1 ms; 0.032 ms).",
        "Pass default@8 after rehearsal: 2026-10-06T16:31:35.610Z to 2026-10-06T16:31:55.620Z (20.0 s), 8,052 searches, 0 failed, p50 19.347 ms, mean 19.873 ms, p99 36.059 ms, 402.4 QPS (1000/QPS 2.485 ms).",
        "Pass exact after default@8: 2026-10-06T16:33:01.638Z to 2026-10-06T16:34:01.644Z (60.0 s), 8,752 searches, 0 failed, p50 6.677 ms, mean 6.852 ms, p99 10.001 ms, 145.9 QPS (1000/QPS 6.856 ms).",
        "Pass default@1 after exact: 2026-10-06T16:34:19.658Z to 2026-10-06T16:34:39.660Z (20.0 s), 3,862 searches, 0 failed, p50 4.957 ms, mean 5.175 ms, p99 7.934 ms, 193.1 QPS (1000/QPS 5.179 ms).",
        "Passes in the order run: default@8, exact, default@1; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 63,069 sent, 0 failed (warmupErrors). Index after the searches: ready, 524 of 524 indexed (1 data part(s), index files on 1 of them, 524 of 524 live rows in indexed parts; 0 merge(s) running, 0 unfinished mutation(s); plan of the default search uses the Skip index vec_idx on 1 of 1 parts).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-clickhouse (1b38cf38a4a4) on network gvb-clickhouse_default at container-ip:8123, not through docker-proxy localhost:8123. Open after the passes: container-ip:8123 x1 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-clickhouse: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-clickhouse: 723 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@8 3492/3492 MHz, average 3492.0/3492.5, performance, outside load 0.23 before the warm-up, 0.21 during, client CPU 1.062 ms per search; exact 3492/3492 MHz, average 3492.0/3492.3, performance, outside load 0.24 before the warm-up, 0.24 during, client CPU 0.889 ms per search; default@1 3492/3492 MHz, average 3492.0/3492.3, performance, outside load 0.24 before the warm-up, 0.24 during, client CPU 0.886 ms per search."
      ]
    },
    {
      "name": "oracle",
      "engine": "Oracle AI Database Free 23.26 (23ai line, VECTOR FLOAT32)",
      "index": "HNSW in-memory neighbor graph NEIGHBORS=16 EFCONSTRUCTION=128, EFSEARCH=100 per query, cosine; exact mode = FETCH EXACT FIRST (full scan); Oracle Free caps itself at 2 CPUs (cpu_count 2 in V$PARAMETER, edition FREE in V$INSTANCE, 8 host CPUs in V$OSSTAT NUM_CPUS; the 2 CPU thread limit is Oracle\u0027s documented Free edition limit)",
      "hosting": "compose",
      "searchSettings": {
        "EFCONSTRUCTION": "128",
        "EFSEARCH": "100",
        "NEIGHBORS": "16"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.2915182,
        "rowsPerSecond": 1797.4864005060404,
        "countMatchSeconds": 0.0500013,
        "indexSeconds": 3.4270415,
        "indexNote": "HNSW graph holds 524 of 524 rows, change log waiting: 0 inserts and 0 deletes, USER_INDEXES status VALID, plan of the default search: VECTOR INDEX HNSW SCAN \u003E TABLE ACCESS BY INDEX ROWID, index used by 0 queries so far; Oracle Free caps itself at 2 CPUs (cpu_count 2 in V$PARAMETER, edition FREE in V$INSTANCE, 8 host CPUs in V$OSSTAT NUM_CPUS; the 2 CPU thread limit is Oracle\u0027s documented Free edition limit); graph rebuilt in 2.2 s"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 21108,
        "p50Ms": 0.92345,
        "p95Ms": 1.0947,
        "p99Ms": 1.799986,
        "qps": {
          "8": 2803.3946138388146,
          "1": 1055.3737634082415
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 23348,
        "exactP50Ms": 2.5026,
        "exactP95Ms": 2.9515649999999996,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "8": 0.6432,
          "1": 0.8864
        }
      },
      "ram": {
        "bytes": 2279553892,
        "text": "2.12 GiB (docker stats)"
      },
      "disk": {
        "bytes": 0,
        "text": "0 B added by this load (6.9 GiB (whole engine data folder), 6.9 GiB before)"
      },
      "passOrder": [
        "default@8",
        "default@1",
        "exact"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "HNSW graph holds 524 of 524 rows, change log waiting: 0 inserts and 0 deletes, USER_INDEXES status VALID, plan of the default search: VECTOR INDEX HNSW SCAN \u003E TABLE ACCESS BY INDEX ROWID, index used by 0 queries so far; Oracle Free caps itself at 2 CPUs (cpu_count 2 in V$PARAMETER, edition FREE in V$INSTANCE, 8 host CPUs in V$OSSTAT NUM_CPUS; the 2 CPU thread limit is Oracle\u0027s documented Free edition limit)"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "HNSW graph holds 524 of 524 rows, change log waiting: 0 inserts and 0 deletes, USER_INDEXES status VALID, plan of the default search: VECTOR INDEX HNSW SCAN \u003E TABLE ACCESS BY INDEX ROWID, index used by 407925 queries so far; Oracle Free caps itself at 2 CPUs (cpu_count 2 in V$PARAMETER, edition FREE in V$INSTANCE, 8 host CPUs in V$OSSTAT NUM_CPUS; the 2 CPU thread limit is Oracle\u0027s documented Free edition limit)"
        }
      },
      "durability": "By these settings a committed row should survive a crash or power loss. The sink uses a plain COMMIT and commit_logging, commit_wait and commit_write are unset in the database, so every commit waits for its redo to be written. Measured 2026-10-04: 50 separate client commits raised V$SYSSTAT \u0027redo synch writes\u0027 by 55 (the 50 commits plus the CREATE and DROP of the probe table), and the log writer and the datafile writer hold their files open with O_DSYNC (open flags 02110002, filesystemio_options none). The database runs NOARCHIVELOG (V$DATABASE.LOG_MODE), so redo serves crash recovery only and there is no point-in-time restore. The HNSW graph lives in the 768 MB vector memory pool (oracle-init/01-vector-memory.sh) and is not the durable copy; the table is. Not tested by cutting power; whether the disk\u0027s own write cache reaches the media was not checked. CPU: Oracle Free caps itself at 2 CPUs (cpu_count 2 in V$PARAMETER, edition FREE in V$INSTANCE, 8 host CPUs in V$OSSTAT NUM_CPUS; the 2 CPU thread limit is Oracle\u0027s documented Free edition limit).",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started oracle.compose.yaml with every container created on CPUs 2-3,6-7: gvb-oracle cpuset 2-3,6-7, gvb-oracle-seed cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [oracle-init/01-vector-memory.sh 6452e4e1953e; oracle.compose.yaml 2a7fd367012f]. This run created the engine from these files.",
        "Index after the load: ready, 524 of 524 indexed (HNSW graph holds 524 of 524 rows, change log waiting: 0 inserts and 0 deletes, USER_INDEXES status VALID, plan of the default search: VECTOR INDEX HNSW SCAN \u003E TABLE ACCESS BY INDEX ROWID, index used by 0 queries so far; Oracle Free caps itself at 2 CPUs (cpu_count 2 in V$PARAMETER, edition FREE in V$INSTANCE, 8 host CPUs in V$OSSTAT NUM_CPUS; the 2 CPU thread limit is Oracle\u0027s documented Free edition limit)). Durability: By these settings a committed row should survive a crash or power loss. The sink uses a plain COMMIT and commit_logging, commit_wait and commit_write are unset in the database, so every commit waits for its redo to be written. Measured 2026-10-04: 50 separate client commits raised V$SYSSTAT \u0027redo synch writes\u0027 by 55 (the 50 commits plus the CREATE and DROP of the probe table), and the log writer and the datafile writer hold their files open with O_DSYNC (open flags 02110002, filesystemio_options none). The database runs NOARCHIVELOG (V$DATABASE.LOG_MODE), so redo serves crash recovery only and there is no point-in-time restore. The HNSW graph lives in the 768 MB vector memory pool (oracle-init/01-vector-memory.sh) and is not the durable copy; the table is. Not tested by cutting power; whether the disk\u0027s own write cache reaches the media was not checked. CPU: Oracle Free caps itself at 2 CPUs (measured 2026-10-04: cpu_count 2 in V$PARAMETER, edition FREE in V$INSTANCE, 8 host CPUs in V$OSSTAT NUM_CPUS; the 2 CPU thread limit is Oracle\u0027s documented Free edition limit; this sink reads the live value when the index state is read)..",
        "Outside load when searching began: 0.24 CPUs on average over the last 6 s and 0.24 over the last few seconds (limit 0.3), a quiet box. The load average 1.85 3.98 3.98 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@8 86,247 searches in 30.0 s, default@1 31,569 searches in 30.0 s, exact 11,662 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@8: warm-up 15.0 s and 46,864 searches (0 failed) at 8 searchers; QPS of the last windows 3,452, 2,436, 3,257 (windows of at least 2 s and 100 searches) (its last 3 windows differed by more than 5%); trial of 7,553 searches in 3.2 s at 8 searchers: 2,390 QPS against the settled 3,257 QPS, 36% apart (limit 10%), so the warm-up was EXTENDED once (30.5 s and 87,862 searches (0 failed) at 8 searchers; QPS of its older 7 windows 2,878 (windows 3,468, 2,412, 3,260, 2,644, 2,826, 3,106, 2,432) and of its newer 7 2,884 (windows 3,458, 2,462, 3,204, 2,798, 2,688, 3,211, 2,367), 0.2% apart (limit 5%) (windows of at least 2 s and 100 searches)); second trial of 9,279 searches in 3.0 s at 8 searchers: 3,092 QPS against the settled 2,884 QPS, 7% apart (limit 10%), inside its newer windows\u0027 range 2,367 to 3,458 QPS; settled after the extension; the timed pass 2,803 QPS, 3% from the settled 2,884 QPS (limit 10%); default@1: warm-up 15.0 s and 15,842 searches (0 failed) at 1 searcher; p50 of the last windows 0.911, 0.936, 0.927 ms (windows of at least 2 s and 100 searches); trial of 3,171 searches in 3.0 s at 1 searcher: p50 0.922 ms against the settled p50 0.927 ms, 0% apart (limit 10%); the timed pass p50 0.923 ms, 0% from the settled p50 0.927 ms (limit 10% or 0.1 ms; 0.003 ms); exact: warm-up 15.0 s and 5,827 searches (0 failed) at 1 searcher; p50 of the last windows 2.516, 2.489, 2.505 ms (windows of at least 2 s and 100 searches); trial of 1,169 searches in 3.0 s at 1 searcher: p50 2.504 ms against the settled p50 2.505 ms, 0% apart (limit 10%); the timed pass p50 2.503 ms, 0% from the settled p50 2.505 ms (limit 10% or 0.1 ms; 0.002 ms).",
        "Pass default@8 after rehearsal: 2026-10-06T16:37:41.563Z to 2026-10-06T16:38:01.647Z (20.1 s), 56,306 searches, 0 failed, p50 1.631 ms, mean 2.851 ms, p99 3.537 ms, 2803.4 QPS (1000/QPS 0.357 ms).",
        "Pass default@1 after default@8: 2026-10-06T16:38:19.669Z to 2026-10-06T16:38:39.670Z (20.0 s), 21,108 searches, 0 failed, p50 0.923 ms, mean 0.946 ms, p99 1.800 ms, 1055.4 QPS (1000/QPS 0.948 ms).",
        "Pass exact after default@1: 2026-10-06T16:38:57.692Z to 2026-10-06T16:39:57.694Z (60.0 s), 23,348 searches, 0 failed, p50 2.503 ms, mean 2.567 ms, p99 3.805 ms, 389.1 QPS (1000/QPS 2.570 ms).",
        "Passes in the order run: default@8, default@1, exact; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 349,169 sent, 0 failed (warmupErrors). Index after the searches: ready, 524 of 524 indexed (HNSW graph holds 524 of 524 rows, change log waiting: 0 inserts and 0 deletes, USER_INDEXES status VALID, plan of the default search: VECTOR INDEX HNSW SCAN \u003E TABLE ACCESS BY INDEX ROWID, index used by 407925 queries so far; Oracle Free caps itself at 2 CPUs (cpu_count 2 in V$PARAMETER, edition FREE in V$INSTANCE, 8 host CPUs in V$OSSTAT NUM_CPUS; the 2 CPU thread limit is Oracle\u0027s documented Free edition limit)).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-oracle (7900917130ab) on network gvb-oracle_default at container-ip:1521, not through docker-proxy localhost:1521. Open after the passes: container-ip:1521 x14 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-oracle: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-oracle: 85 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@8 3492/3492 MHz, average 3167.8/3213.8, performance, outside load 0.24 before the warm-up, 0.24 during, client CPU 0.643 ms per search; default@1 3492/3492 MHz, average 3492.4/3492.1, performance, outside load 0.24 before the warm-up, 0.22 during, client CPU 0.886 ms per search; exact 3492/3492 MHz, average 3492.2/3491.6, performance, outside load 0.23 before the warm-up, 0.21 during, client CPU 0.949 ms per search.",
        "WARNING: clock off its pinned value during oracle default@8: the engine CPUs averaged 3168 MHz, -9.5% against the pinned 3500 MHz; the client CPUs averaged 3214 MHz, -8.2% against the pinned 3500 MHz (limit 1%); this pass is not comparable with passes at the pinned clock."
      ]
    },
    {
      "name": "sql",
      "engine": "Microsoft SQL Server 2025 (RTM-CU9) (KB5122048) - 17.0.5005.3 (X64), Enterprise Developer Edition (64-bit), in container gvb-mssql (image mcr.microsoft.com/mssql/server:2025-CU9-ubuntu-24.04@sha256:2b5b581621126574f3d1f75e78d3eebe8d05aedb59ad0cfdf9aa42cb0634d726; cpuset 2-3,6-7 from its creation; SQL Server counts 4 CPU(s) and runs 4 visible scheduler(s), affinity AUTO)",
      "index": "exact VECTOR_DISTANCE cosine, no vector index (full scan)",
      "hosting": "compose",
      "searchSettings": {
        "description": "exact VECTOR_DISTANCE cosine, no vector index (full scan)"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.8540889,
        "rowsPerSecond": 613.5192718228747,
        "countMatchSeconds": 0.004144
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 4821,
        "p50Ms": 4.0501,
        "p95Ms": 4.6013,
        "p99Ms": 6.079480000000004,
        "qps": {
          "1": 241.0135057349616,
          "8": 613.6218902390702
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 0,
        "clientCpuMsPerSearch": {
          "1": 1.2084,
          "8": 1.1921
        }
      },
      "ram": {
        "bytes": 555011276,
        "text": "529.3 MiB (docker stats)"
      },
      "disk": {
        "bytes": 5398528,
        "text": "5.15 MiB (table and its indexes, reserved pages)"
      },
      "passOrder": [
        "default@1",
        "default@8"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 0,
          "totalVectors": 524,
          "detail": "no index used, exact scan by design: dbo.gvb_gvbbench_eshoponweb in GvbBench holds 524 rows; sys.vector_indexes lists no index on the table; indexes: PK_gvb_gvbbench_eshoponweb CLUSTERED, IX_gvb_gvbbench_eshoponweb_DocKey NONCLUSTERED; a real search under SET STATISTICS XML ON ran: Clustered Index Scan of GvbBench.gvb_gvbbench_eshoponweb through PK_gvb_gvbbench_eshoponweb (IndexKind Clustered), 524 rows read, no vector index operator, 10 hits"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 0,
          "totalVectors": 524,
          "detail": "no index used, exact scan by design: dbo.gvb_gvbbench_eshoponweb in GvbBench holds 524 rows; sys.vector_indexes lists no index on the table; indexes: PK_gvb_gvbbench_eshoponweb CLUSTERED, IX_gvb_gvbbench_eshoponweb_DocKey NONCLUSTERED; a real search under SET STATISTICS XML ON ran: Clustered Index Scan of GvbBench.gvb_gvbbench_eshoponweb through PK_gvb_gvbbench_eshoponweb (IndexKind Clustered), 524 rows read, no vector index operator, 10 hits"
        }
      },
      "durability": "A commit returns after its transaction-log records are written to disk (SQL Server write-ahead logging; delayed durability is DISABLED); a database created here copies the model database: recovery model FULL, page_verify CHECKSUM; no global trace flags are enabled; mssql.conf of container gvb-mssql (/var/opt/mssql/mssql.conf, on the host at ~/gvb-data/engines/mssql/mssql.conf) does not exist, so it sets: nothing; it has no [control] or [traceflag] entry, so SQL Server\u0027s own Linux defaults for flushing writes apply (Microsoft\u0027s Linux performance guide names trace flag 3982 as that default; read from the guide, not tested here). Not tested by cutting power; whether the disk\u0027s own write cache reaches the media was not checked. Container settings from its environment (names only): MSSQL_AGENT_ENABLED, MSSQL_MEMORY_LIMIT_MB, MSSQL_PID, MSSQL_RPC_PORT, MSSQL_SA_PASSWORD.",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started mssql.compose.yaml with every container created on CPUs 2-3,6-7: gvb-mssql cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [mssql.compose.yaml b7359663cc15]. This run created the engine from these files.",
        "Index after the load: ready, 0 of 524 indexed (no index used, exact scan by design: dbo.gvb_gvbbench_eshoponweb in GvbBench holds 524 rows; sys.vector_indexes lists no index on the table; indexes: PK_gvb_gvbbench_eshoponweb CLUSTERED, IX_gvb_gvbbench_eshoponweb_DocKey NONCLUSTERED; a real search under SET STATISTICS XML ON ran: Clustered Index Scan of GvbBench.gvb_gvbbench_eshoponweb through PK_gvb_gvbbench_eshoponweb (IndexKind Clustered), 524 rows read, no vector index operator, 10 hits). Durability: A commit returns after its transaction-log records are written to disk (SQL Server write-ahead logging; delayed durability is DISABLED); a database created here copies the model database: recovery model FULL, page_verify CHECKSUM; no global trace flags are enabled; mssql.conf of container gvb-mssql (/var/opt/mssql/mssql.conf, on the host at ~/gvb-data/engines/mssql/mssql.conf) does not exist, so it sets: nothing; it has no [control] or [traceflag] entry, so SQL Server\u0027s own Linux defaults for flushing writes apply (Microsoft\u0027s Linux performance guide names trace flag 3982 as that default; read from the guide, not tested here). Not tested by cutting power; whether the disk\u0027s own write cache reaches the media was not checked. Container settings from its environment (names only): MSSQL_AGENT_ENABLED, MSSQL_MEMORY_LIMIT_MB, MSSQL_PID, MSSQL_RPC_PORT, MSSQL_SA_PASSWORD..",
        "Outside load when searching began: 0.21 CPUs on average over the last 3 s and 0.21 over the last few seconds (limit 0.3), a quiet box. The load average 1.96 3.26 3.70 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@1 7,314 searches in 30.0 s, default@8 18,147 searches in 30.1 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@1: warm-up 15.0 s and 3,639 searches (0 failed) at 1 searcher; p50 of the last windows 4.027, 4.039, 4.051 ms (windows of at least 2 s and 100 searches); trial of 724 searches in 3.0 s at 1 searcher: p50 4.053 ms against the settled p50 4.039 ms, 0% apart (limit 10%); the timed pass p50 4.050 ms, 0% from the settled p50 4.039 ms (limit 10% or 0.1 ms; 0.011 ms); default@8: warm-up 15.0 s and 8,671 searches (0 failed) at 8 searchers; QPS of the last windows 555, 574, 582 (windows of at least 2 s and 100 searches); trial of 1,866 searches in 3.0 s at 8 searchers: 621 QPS against the settled 574 QPS, 8% apart (limit 10%); the timed pass 614 QPS, 7% from the settled 574 QPS (limit 10%).",
        "Pass default@1 after rehearsal: 2026-10-06T16:41:30.293Z to 2026-10-06T16:41:50.296Z (20.0 s), 4,821 searches, 0 failed, p50 4.050 ms, mean 4.145 ms, p99 6.079 ms, 241.0 QPS (1000/QPS 4.149 ms).",
        "Pass default@8 after default@1: 2026-10-06T16:42:08.314Z to 2026-10-06T16:42:28.325Z (20.0 s), 12,279 searches, 0 failed, p50 12.538 ms, mean 13.031 ms, p99 19.818 ms, 613.6 QPS (1000/QPS 1.630 ms).",
        "Passes in the order run: default@1, default@8; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 40,361 sent, 0 failed (warmupErrors). Index after the searches: ready, 0 of 524 indexed (no index used, exact scan by design: dbo.gvb_gvbbench_eshoponweb in GvbBench holds 524 rows; sys.vector_indexes lists no index on the table; indexes: PK_gvb_gvbbench_eshoponweb CLUSTERED, IX_gvb_gvbbench_eshoponweb_DocKey NONCLUSTERED; a real search under SET STATISTICS XML ON ran: Clustered Index Scan of GvbBench.gvb_gvbbench_eshoponweb through PK_gvb_gvbbench_eshoponweb (IndexKind Clustered), 524 rows read, no vector index operator, 10 hits).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-mssql (703c36ce4875) on network gvb-mssql_default at container-ip:1433, not through docker-proxy localhost:14330. Open after the passes: container-ip:1433 x9 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-mssql: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-mssql: 86 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@1 3492/3492 MHz, average 3492.4/3492.4, performance, outside load 0.22 before the warm-up, 0.20 during, client CPU 1.208 ms per search; default@8 3492/3492 MHz, average 3492.0/3492.2, performance, outside load 0.22 before the warm-up, 0.24 during, client CPU 1.192 ms per search."
      ]
    },
    {
      "name": "opensearch",
      "engine": "OpenSearch 3.9.0 (k-NN plugin, faiss)",
      "index": "faiss HNSW float32, no compression, m=16, ef_construction=128, cosinesimil; search k=top, ef_search=100; 1 shard, 0 replicas; graph built at any segment size (approximate_threshold=0); force-merged to one segment after the load",
      "hosting": "compose",
      "searchSettings": {
        "approximate_threshold": "0",
        "ef_construction": "128",
        "ef_search": "100",
        "k": "top",
        "m": "16"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 1.6279591,
        "rowsPerSecond": 321.8754082949627,
        "countMatchSeconds": 0.1637854,
        "indexSeconds": 0.8990546,
        "indexNote": "force-merged to one segment, graph ready after 0.9 s: every segment searched through its HNSW graph, covering 524 of 524 vectors: 1 segment(s), 0 merge(s) running, profiled probe search: ann_search_count 1 (segments tried through a graph), exact_search_count 0 (of those, scanned instead), approximate_threshold 0; force-merged to one segment on purpose, so one graph answers every search"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 9675,
        "p50Ms": 2.027,
        "p95Ms": 2.31532,
        "p99Ms": 2.7935820000000002,
        "qps": {
          "8": 1476.537174863042,
          "1": 483.70290668500513
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 22550,
        "exactP50Ms": 2.6149,
        "exactP95Ms": 2.9371549999999997,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "8": 0.5104,
          "1": 0.465
        }
      },
      "ram": {
        "bytes": 2734820425,
        "text": "2.55 GiB (docker stats)"
      },
      "disk": {
        "bytes": 10013916,
        "text": "9.55 MiB added by this load (12.86 MiB (whole engine data folder), 3.31 MiB before)"
      },
      "passOrder": [
        "default@8",
        "exact",
        "default@1"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "every segment searched through its HNSW graph, covering 524 of 524 vectors: 1 segment(s), 0 merge(s) running, profiled probe search: ann_search_count 1 (segments tried through a graph), exact_search_count 0 (of those, scanned instead), approximate_threshold 0; force-merged to one segment on purpose, so one graph answers every search"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "every segment searched through its HNSW graph, covering 524 of 524 vectors: 1 segment(s), 0 merge(s) running, profiled probe search: ann_search_count 1 (segments tried through a graph), exact_search_count 0 (of those, scanned instead), approximate_threshold 0; force-merged to one segment on purpose, so one graph answers every search"
        }
      },
      "durability": "Every acknowledged bulk request is fsynced to the translog before the answer: index.translog.durability=request, the OpenSearch default, which neither this sink nor opensearch.compose.yaml overrides (OpenSearchReadinessTests reads it back from the live index). By that setting a process crash or power loss loses no acknowledged write; this is read from the setting, not shown by pulling power. One node and no replicas, so a lost disk loses the data.",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started opensearch.compose.yaml with every container created on CPUs 2-3,6-7: gvb-opensearch cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [opensearch.compose.yaml 028b44c40129]. This run created the engine from these files.",
        "Index after the load: ready, 524 of 524 indexed (every segment searched through its HNSW graph, covering 524 of 524 vectors: 1 segment(s), 0 merge(s) running, profiled probe search: ann_search_count 1 (segments tried through a graph), exact_search_count 0 (of those, scanned instead), approximate_threshold 0; force-merged to one segment on purpose, so one graph answers every search). Durability: Every acknowledged bulk request is fsynced to the translog before the answer: index.translog.durability=request, the OpenSearch default, which neither this sink nor opensearch.compose.yaml overrides (OpenSearchReadinessTests reads it back from the live index). By that setting a process crash or power loss loses no acknowledged write; this is read from the setting, not shown by pulling power. One node and no replicas, so a lost disk loses the data..",
        "WARNING: processes outside the benchmark used 0.62 CPUs on average over the last 3 s and 0.62 over the last few seconds when searching began (limit 0.3): the box was busy as the untimed rehearsal started. Each timed pass still waits for a quiet box before its warm-up, and one that ran busy is flagged \u0027busy box\u0027 on its own. The load average 3.77 3.37 3.65 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@8 26,230 searches in 30.0 s, exact 10,863 searches in 30.0 s, default@1 14,452 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@8: warm-up 15.0 s and 21,173 searches (0 failed) at 8 searchers; QPS of the last windows 1,482, 1,476, 1,484 (windows of at least 2 s and 100 searches); trial of 4,418 searches in 3.0 s at 8 searchers: 1,470 QPS against the settled 1,482 QPS, 1% apart (limit 10%); the timed pass 1,477 QPS, 0% from the settled 1,482 QPS (limit 10%); exact: warm-up 15.0 s and 5,610 searches (0 failed) at 1 searcher; p50 of the last windows 2.639, 2.640, 2.628 ms (windows of at least 2 s and 100 searches); trial of 1,111 searches in 3.0 s at 1 searcher: p50 2.632 ms against the settled p50 2.639 ms, 0% apart (limit 10%); the timed pass p50 2.615 ms, 1% from the settled p50 2.639 ms (limit 10% or 0.1 ms; 0.024 ms); default@1: warm-up 15.0 s and 7,242 searches (0 failed) at 1 searcher; p50 of the last windows 2.025, 2.031, 2.025 ms (windows of at least 2 s and 100 searches); trial of 1,441 searches in 3.0 s at 1 searcher: p50 2.025 ms against the settled p50 2.025 ms, 0% apart (limit 10%); the timed pass p50 2.027 ms, 0% from the settled p50 2.025 ms (limit 10% or 0.1 ms; 0.002 ms).",
        "Pass default@8 after rehearsal: 2026-10-06T16:44:48.389Z to 2026-10-06T16:45:08.391Z (20.0 s), 29,534 searches, 0 failed, p50 5.183 ms, mean 5.415 ms, p99 10.628 ms, 1476.5 QPS (1000/QPS 0.677 ms).",
        "Pass exact after default@8: 2026-10-06T16:45:26.403Z to 2026-10-06T16:46:26.405Z (60.0 s), 22,550 searches, 0 failed, p50 2.615 ms, mean 2.658 ms, p99 3.515 ms, 375.8 QPS (1000/QPS 2.661 ms).",
        "Pass default@1 after exact: 2026-10-06T16:46:44.426Z to 2026-10-06T16:47:04.428Z (20.0 s), 9,675 searches, 0 failed, p50 2.027 ms, mean 2.065 ms, p99 2.794 ms, 483.7 QPS (1000/QPS 2.067 ms).",
        "Passes in the order run: default@8, exact, default@1; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 92,540 sent, 0 failed (warmupErrors). Index after the searches: ready, 524 of 524 indexed (every segment searched through its HNSW graph, covering 524 of 524 vectors: 1 segment(s), 0 merge(s) running, profiled probe search: ann_search_count 1 (segments tried through a graph), exact_search_count 0 (of those, scanned instead), approximate_threshold 0; force-merged to one segment on purpose, so one graph answers every search).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-opensearch (aa945bc1d1e9) on network engines_default at container-ip:9200, not through docker-proxy localhost:9201. Open after the passes: container-ip:9200 x1 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-opensearch: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-opensearch: 55 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@8 3492/3492 MHz, average 3492.0/3492.2, performance, outside load 0.21 before the warm-up, 0.20 during, client CPU 0.510 ms per search; exact 3492/3492 MHz, average 3492.1/3492.6, performance, outside load 0.20 before the warm-up, 0.21 during, client CPU 0.472 ms per search; default@1 3492/3492 MHz, average 3492.0/3492.2, performance, outside load 0.21 before the warm-up, 0.21 during, client CPU 0.465 ms per search."
      ]
    },
    {
      "name": "redis",
      "engine": "Redis 8.10.2 (query engine, HASH \u002B vector index)",
      "index": "HNSW TYPE FLOAT32 M=16 EF_CONSTRUCTION=128, EF_RUNTIME=100 per query, cosine; exact mode = FLAT index built on first exact query",
      "hosting": "compose",
      "searchSettings": {
        "EF_CONSTRUCTION": "128",
        "EF_RUNTIME": "100",
        "M": "16"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.0435549,
        "rowsPerSecond": 12030.793320613753,
        "countMatchSeconds": 1.020824,
        "indexSeconds": 0.0254644,
        "indexNote": "HNSW index finished after 0.0 s of waiting: FT.INFO indexing 0, percent_indexed 1, hash_indexing_failures 0, flat_buffer_size 0, num_docs 524, hashes counted with SCAN under gvb_gvbbench_eshoponweb: 524"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 50378,
        "p50Ms": 0.3985,
        "p95Ms": 0.4891,
        "p99Ms": 0.7036379999999975,
        "qps": {
          "8": 5037.759447698817,
          "1": 2518.8969647291574
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 160015,
        "exactP50Ms": 0.3767,
        "exactP95Ms": 0.471,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "8": 0.3691,
          "1": 0.5686
        }
      },
      "ram": {
        "bytes": 266338304,
        "text": "254 MiB (docker stats)"
      },
      "disk": {
        "bytes": 2547239,
        "text": "2.43 MiB added by this load (96.65 MiB (whole engine data folder), 94.22 MiB before)"
      },
      "passOrder": [
        "default@8",
        "exact",
        "default@1"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "FT.INFO indexing 0, percent_indexed 1, hash_indexing_failures 0, flat_buffer_size 0, num_docs 524, hashes counted with SCAN under gvb_gvbbench_eshoponweb: 524"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "FT.INFO indexing 0, percent_indexed 1, hash_indexing_failures 0, flat_buffer_size 0, num_docs 524, hashes counted with SCAN under gvb_gvbbench_eshoponweb: 524"
        }
      },
      "durability": "save \u0022300 1\u0022 and appendonly no (redis.compose.yaml): an RDB snapshot is written every 5 minutes if at least one key changed, and on a clean stop; there is no append-only log, so a crash loses every write since the last snapshot (up to 5 minutes plus the time a snapshot takes; measured with docker kill, which is SIGKILL, right after a 2,000-vector load: none of the 2,000 hashes were there after the restart, and the restart loaded an older snapshot that still held keys of a collection that had been dropped since)",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started redis.compose.yaml with every container created on CPUs 2-3,6-7: gvb-redis cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [redis.compose.yaml 7448634e0d92]. This run created the engine from these files.",
        "Index after the load: ready, 524 of 524 indexed (FT.INFO indexing 0, percent_indexed 1, hash_indexing_failures 0, flat_buffer_size 0, num_docs 524, hashes counted with SCAN under gvb_gvbbench_eshoponweb: 524). Durability: save \u0022300 1\u0022 and appendonly no (redis.compose.yaml): an RDB snapshot is written every 5 minutes if at least one key changed, and on a clean stop; there is no append-only log, so a crash loses every write since the last snapshot (up to 5 minutes plus the time a snapshot takes; measured with docker kill, which is SIGKILL, right after a 2,000-vector load: none of the 2,000 hashes were there after the restart, and the restart loaded an older snapshot that still held keys of a collection that had been dropped since).",
        "Outside load when searching began: 0.17 CPUs on average over the last 1 s and 0.17 over the last few seconds (limit 0.3), a quiet box. The load average 1.81 3.04 3.51 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@8 150,346 searches in 30.0 s, exact 77,412 searches in 30.0 s, default@1 75,345 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@8: warm-up 15.0 s and 75,594 searches (0 failed) at 8 searchers; QPS of the last windows 5,120, 5,048, 4,864 (windows of at least 2 s and 100 searches) (its last 3 windows differed by more than 5%); trial of 15,388 searches in 3.0 s at 8 searchers: 5,127 QPS against the settled 5,048 QPS, 2% apart (limit 10%), so the warm-up was EXTENDED once (30.0 s and 151,551 searches (0 failed) at 8 searchers; QPS of its older 6 windows 5,043 (windows 4,982, 5,092, 5,022, 5,076, 5,109, 4,976) and of its newer 6 5,075 (windows 5,121, 5,052, 5,046, 4,994, 5,065, 5,175), 0.6% apart (limit 5%) (windows of at least 2 s and 100 searches)); second trial of 15,561 searches in 3.0 s at 8 searchers: 5,184 QPS against the settled 5,075 QPS, 2% apart (limit 10%), outside its newer windows\u0027 range 4,994 to 5,175 QPS; settled after the extension; the timed pass 5,038 QPS, 1% from the settled 5,075 QPS (limit 10%); exact: warm-up 15.0 s and 40,000 searches (0 failed) at 1 searcher; p50 of the last windows 0.363, 0.344, 0.367 ms (windows of at least 2 s and 100 searches) (its last 3 windows differed by more than 5%); trial of 7,985 searches in 3.0 s at 1 searcher: p50 0.380 ms against the settled p50 0.363 ms, 5% apart (limit 10%), so the warm-up was EXTENDED once (30.0 s and 79,780 searches (0 failed) at 1 searcher; p50 of its older 6 windows 0.377 ms (windows 0.373, 0.387, 0.379, 0.373, 0.363, 0.386) and of its newer 6 0.382 ms (windows 0.365, 0.387, 0.377, 0.388, 0.386, 0.387), 1.2% apart (limit 5% or 0.1 ms; 0.005 ms) (windows of at least 2 s and 100 searches)); second trial of 8,129 searches in 3.0 s at 1 searcher: p50 0.361 ms against the settled p50 0.382 ms, 6% apart (limit 10%), outside its newer windows\u0027 range 0.365 to 0.388 ms; settled after the extension; the timed pass p50 0.377 ms, 1% from the settled p50 0.382 ms (limit 10% or 0.1 ms; 0.005 ms); default@1: warm-up 15.0 s and 37,930 searches (0 failed) at 1 searcher; p50 of the last windows 0.410, 0.405, 0.370 ms (windows of at least 2 s and 100 searches) (its last 3 windows differed by more than 5%); trial of 7,551 searches in 3.0 s at 1 searcher: p50 0.403 ms against the settled p50 0.405 ms, 1% apart (limit 10%), so the warm-up was EXTENDED once (30.0 s and 75,942 searches (0 failed) at 1 searcher; p50 of its older 6 windows 0.398 ms (windows 0.403, 0.405, 0.387, 0.395, 0.388, 0.407) and of its newer 6 0.390 ms (windows 0.394, 0.379, 0.393, 0.397, 0.393, 0.385), 2.0% apart (limit 5% or 0.1 ms; 0.008 ms) (windows of at least 2 s and 100 searches)); second trial of 7,583 searches in 3.0 s at 1 searcher: p50 0.399 ms against the settled p50 0.390 ms, 2% apart (limit 10%), outside its newer windows\u0027 range 0.379 to 0.397 ms; settled after the extension; the timed pass p50 0.399 ms, 2% from the settled p50 0.390 ms (limit 10% or 0.1 ms; 0.008 ms).",
        "Pass default@8 after rehearsal: 2026-10-06T16:50:00.441Z to 2026-10-06T16:50:20.443Z (20.0 s), 100,763 searches, 0 failed, p50 1.606 ms, mean 1.586 ms, p99 2.248 ms, 5037.8 QPS (1000/QPS 0.199 ms).",
        "Pass exact after default@8: 2026-10-06T16:51:26.498Z to 2026-10-06T16:52:26.498Z (60.0 s), 160,015 searches, 0 failed, p50 0.377 ms, mean 0.373 ms, p99 0.690 ms, 2666.9 QPS (1000/QPS 0.375 ms).",
        "Pass default@1 after exact: 2026-10-06T16:53:32.683Z to 2026-10-06T16:53:52.683Z (20.0 s), 50,378 searches, 0 failed, p50 0.399 ms, mean 0.395 ms, p99 0.704 ms, 2518.9 QPS (1000/QPS 0.397 ms).",
        "Passes in the order run: default@8, exact, default@1; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 979,656 sent, 0 failed (warmupErrors). Index after the searches: ready, 524 of 524 indexed (FT.INFO indexing 0, percent_indexed 1, hash_indexing_failures 0, flat_buffer_size 0, num_docs 524, hashes counted with SCAN under gvb_gvbbench_eshoponweb: 524).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-redis (78e85925c519) on network gvb-redis_default at container-ip:6379, not through docker-proxy localhost:6379. Open after the passes: container-ip:6379 x1 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-redis: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-redis: 16 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@8 3492/3492 MHz, average 3492.2/3492.0, performance, outside load 0.20 before the warm-up, 0.20 during, client CPU 0.369 ms per search; exact 3492/3492 MHz, average 3492.1/3492.0, performance, outside load 0.19 before the warm-up, 0.20 during, client CPU 0.567 ms per search; default@1 3492/3492 MHz, average 3492.1/3491.7, performance, outside load 0.19 before the warm-up, 0.19 during, client CPU 0.569 ms per search."
      ]
    },
    {
      "name": "chroma",
      "engine": "Chroma 1.4.4 (single node, REST API v2)",
      "index": "HNSW M=16 ef_construction=128, ef_search=100 (Chroma default), cosine; approximate only (no exact mode)",
      "hosting": "compose",
      "searchSettings": {
        "M": "16",
        "ef_construction": "128",
        "ef_search": "100"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.6071146,
        "rowsPerSecond": 863.0989931719646,
        "countMatchSeconds": 0.0222435,
        "indexSeconds": 0.0181165,
        "indexNote": "Chroma has no index build to wait for and exposes no index status; checked in 0.0 s: Chroma exposes no index status (indexing_status endpoint: HTTP 500 {\u0022error\u0022:\u0022InternalError\u0022,\u0022message\u0022:\u0022Method scout_logs is not implemented\u0022}), so indexed vectors are not reported and ready only means the two counts agree to within 1 percent: count endpoint 524, a vector search asking for 524 results returned 524"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 9377,
        "p50Ms": 2.106,
        "p95Ms": 2.3915399999999987,
        "p99Ms": 2.629172,
        "qps": {
          "8": 777.109955365809,
          "1": 468.8430259599889
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 0,
        "clientCpuMsPerSearch": {
          "8": 0.4719,
          "1": 0.4764
        }
      },
      "ram": {
        "bytes": 48402268,
        "text": "46.16 MiB (docker stats)"
      },
      "disk": {
        "bytes": 424100,
        "text": "414.16 KiB added by this load (2.01 GiB (whole engine data folder), 2.01 GiB before)"
      },
      "passOrder": [
        "default@8",
        "default@1"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "totalVectors": 524,
          "detail": "Chroma exposes no index status (indexing_status endpoint: HTTP 500 {\u0022error\u0022:\u0022InternalError\u0022,\u0022message\u0022:\u0022Method scout_logs is not implemented\u0022}), so indexed vectors are not reported and ready only means the two counts agree to within 1 percent: count endpoint 524, a vector search asking for 524 results returned 524"
        },
        "afterSearch": {
          "ready": true,
          "totalVectors": 524,
          "detail": "Chroma exposes no index status (indexing_status endpoint: HTTP 500 {\u0022error\u0022:\u0022InternalError\u0022,\u0022message\u0022:\u0022Method scout_logs is not implemented\u0022}), so indexed vectors are not reported and ready only means the two counts agree to within 1 percent: count endpoint 524, a vector search asking for 524 results returned 524"
        }
      },
      "durability": "SQLite rollback journal with its default synchronous=FULL under the data directory (chroma.compose.yaml sets IS_PERSISTENT=1 and PERSIST_DIRECTORY=/data and no sync setting): every write is committed to chroma.sqlite3 with fsync of the journal, the directory and the database file before the call returns (strace: 15 database fsyncs and 38 journal fsyncs for one create, four 500-vector upserts and one delete), and the HNSW files are written every sync_threshold=1000 vectors and rebuilt from the SQLite log after a crash; measured with kill -9 and a restart: all 50, 300 and 2,000 vectors were still stored and searchable",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started chroma.compose.yaml with every container created on CPUs 2-3,6-7: gvb-chroma cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [chroma.compose.yaml f7ed9d3c8ff7]. This run created the engine from these files.",
        "Index after the load: ready, ? of 524 indexed (Chroma exposes no index status (indexing_status endpoint: HTTP 500 {\u0022error\u0022:\u0022InternalError\u0022,\u0022message\u0022:\u0022Method scout_logs is not implemented\u0022}), so indexed vectors are not reported and ready only means the two counts agree to within 1 percent: count endpoint 524, a vector search asking for 524 results returned 524). Durability: SQLite rollback journal with its default synchronous=FULL under the data directory (chroma.compose.yaml sets IS_PERSISTENT=1 and PERSIST_DIRECTORY=/data and no sync setting): every write is committed to chroma.sqlite3 with fsync of the journal, the directory and the database file before the call returns (strace: 15 database fsyncs and 38 journal fsyncs for one create, four 500-vector upserts and one delete), and the HNSW files are written every sync_threshold=1000 vectors and rebuilt from the SQLite log after a crash; measured with kill -9 and a restart: all 50, 300 and 2,000 vectors were still stored and searchable.",
        "Outside load when searching began: 0.07 CPUs on average over the last 1 s and 0.07 over the last few seconds (limit 0.3), a quiet box. The load average 2.07 2.69 3.21 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@8 23,210 searches in 30.0 s, default@1 14,036 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@8: warm-up 15.0 s and 11,646 searches (0 failed) at 8 searchers; QPS of the last windows 776, 780, 777 (windows of at least 2 s and 100 searches); trial of 2,346 searches in 3.0 s at 8 searchers: 780 QPS against the settled 777 QPS, 0% apart (limit 10%); the timed pass 777 QPS, 0% from the settled 777 QPS (limit 10%); default@1: warm-up 15.0 s and 7,032 searches (0 failed) at 1 searcher; p50 of the last windows 2.114, 2.116, 2.105 ms (windows of at least 2 s and 100 searches); trial of 1,414 searches in 3.0 s at 1 searcher: p50 2.100 ms against the settled p50 2.114 ms, 1% apart (limit 10%); the timed pass p50 2.106 ms, 0% from the settled p50 2.114 ms (limit 10% or 0.1 ms; 0.008 ms).",
        "Pass default@8 after rehearsal: 2026-10-06T16:55:19.991Z to 2026-10-06T16:55:39.998Z (20.0 s), 15,548 searches, 0 failed, p50 10.235 ms, mean 10.291 ms, p99 11.877 ms, 777.1 QPS (1000/QPS 1.287 ms).",
        "Pass default@1 after default@8: 2026-10-06T16:55:58.007Z to 2026-10-06T16:56:18.007Z (20.0 s), 9,377 searches, 0 failed, p50 2.106 ms, mean 2.131 ms, p99 2.629 ms, 468.8 QPS (1000/QPS 2.133 ms).",
        "Passes in the order run: default@8, default@1; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 59,684 sent, 0 failed (warmupErrors). Index after the searches: ready, ? of 524 indexed (Chroma exposes no index status (indexing_status endpoint: HTTP 500 {\u0022error\u0022:\u0022InternalError\u0022,\u0022message\u0022:\u0022Method scout_logs is not implemented\u0022}), so indexed vectors are not reported and ready only means the two counts agree to within 1 percent: count endpoint 524, a vector search asking for 524 results returned 524).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-chroma (dd4da1e51118) on network gvb-chroma_default at container-ip:8000, not through docker-proxy localhost:8000. Open after the passes: container-ip:8000 x8 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-chroma: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-chroma: 9 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@8 3492/3492 MHz, average 3492.0/3492.2, performance, outside load 0.16 before the warm-up, 0.15 during, client CPU 0.472 ms per search; default@1 3492/3492 MHz, average 3492.0/3492.5, performance, outside load 0.15 before the warm-up, 0.15 during, client CPU 0.476 ms per search."
      ]
    },
    {
      "name": "typesense",
      "engine": "Typesense 30.2 (vector search)",
      "index": "HNSW float32 (hnswlib), m=16, ef_construction=128, cosine; search k=top, ef=100; exact mode = filter ordinal:\u003E=0 with flat_search_cutoff; index held in memory",
      "hosting": "compose",
      "searchSettings": {
        "ef": "100",
        "ef_construction": "128",
        "k": "top",
        "m": "16"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.5346849,
        "rowsPerSecond": 980.0164545510823,
        "countMatchSeconds": 0.0019548,
        "indexSeconds": 0.0137439,
        "indexNote": "nothing to build, Typesense inserts into the HNSW graph while writing; ready after 0.0 s: HNSW graph returns 524 of 524 vectors, no writes queued: num_documents 524, pending_write_batches 0, health ok, unfiltered vector query with k above the count returned 524; the graph is built in memory during each write, so there is no later build step"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 4716,
        "p50Ms": 4.1859,
        "p95Ms": 4.539325,
        "p99Ms": 4.965175000000001,
        "qps": {
          "1": 235.77537797727783,
          "8": 576.1259043070017
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 10446,
        "exactP50Ms": 5.66805,
        "exactP95Ms": 6.054525,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "1": 0.5208,
          "8": 0.5918
        }
      },
      "ram": {
        "bytes": 127926272,
        "text": "122 MiB (docker stats)"
      },
      "disk": {
        "bytes": 8554105,
        "text": "8.16 MiB added by this load (279.8 MiB (whole engine data folder), 271.64 MiB before)"
      },
      "passOrder": [
        "exact",
        "default@1",
        "default@8"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "HNSW graph returns 524 of 524 vectors, no writes queued: num_documents 524, pending_write_batches 0, health ok, unfiltered vector query with k above the count returned 524; the graph is built in memory during each write, so there is no later build step"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "HNSW graph returns 524 of 524 vectors, no writes queued: num_documents 524, pending_write_batches 0, health ok, unfiltered vector query with k above the count returned 524; the graph is built in memory during each write, so there is no later build step"
        }
      },
      "durability": "Every acknowledged write is appended to Typesense\u0027s raft log and fsynced before the answer: braft raft_sync=true with raft_sync_policy=0 (sync immediately), read from the running container\u0027s brpc /flags page on 2026-10-04. A restart replays the log from the last snapshot (typesense.compose.yaml sets TYPESENSE_SNAPSHOT_INTERVAL_SECONDS=300) and rebuilds the in-memory HNSW graph, so by those settings a crash loses no acknowledged write, but the restart is slow after a big load. This is read from the settings, not shown by pulling power. One node and no replicas, so a lost disk loses the data.",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started typesense.compose.yaml with every container created on CPUs 2-3,6-7: gvb-typesense cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [typesense.compose.yaml 96a165f62566]. This run created the engine from these files.",
        "Index after the load: ready, 524 of 524 indexed (HNSW graph returns 524 of 524 vectors, no writes queued: num_documents 524, pending_write_batches 0, health ok, unfiltered vector query with k above the count returned 524; the graph is built in memory during each write, so there is no later build step). Durability: Every acknowledged write is appended to Typesense\u0027s raft log and fsynced before the answer: braft raft_sync=true with raft_sync_policy=0 (sync immediately), read from the running container\u0027s brpc /flags page on 2026-10-04. A restart replays the log from the last snapshot (typesense.compose.yaml sets TYPESENSE_SNAPSHOT_INTERVAL_SECONDS=300) and rebuilds the in-memory HNSW graph, so by those settings a crash loses no acknowledged write, but the restart is slow after a big load. This is read from the settings, not shown by pulling power. One node and no replicas, so a lost disk loses the data..",
        "Outside load when searching began: 0.25 CPUs on average over the last 6 s and 0.24 over the last few seconds (limit 0.3), a quiet box. The load average 2.73 2.63 3.10 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: exact 5,238 searches in 30.0 s, default@1 7,052 searches in 30.0 s, default@8 17,273 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. exact: warm-up 15.0 s and 2,621 searches (0 failed) at 1 searcher; p50 of the last windows 5.649, 5.655, 5.664 ms (windows of at least 2 s and 100 searches); trial of 522 searches in 3.0 s at 1 searcher: p50 5.665 ms against the settled p50 5.655 ms, 0% apart (limit 10%); the timed pass p50 5.668 ms, 0% from the settled p50 5.655 ms (limit 10% or 0.1 ms; 0.013 ms); default@1: warm-up 15.0 s and 3,534 searches (0 failed) at 1 searcher; p50 of the last windows 4.195, 4.187, 4.182 ms (windows of at least 2 s and 100 searches); trial of 706 searches in 3.0 s at 1 searcher: p50 4.187 ms against the settled p50 4.187 ms, 0% apart (limit 10%); the timed pass p50 4.186 ms, 0% from the settled p50 4.187 ms (limit 10% or 0.1 ms; 0.002 ms); default@8: warm-up 15.0 s and 8,598 searches (0 failed) at 8 searchers; QPS of the last windows 570, 571, 575 (windows of at least 2 s and 100 searches); trial of 1,729 searches in 3.0 s at 8 searchers: 575 QPS against the settled 571 QPS, 1% apart (limit 10%); the timed pass 576 QPS, 1% from the settled 571 QPS (limit 10%).",
        "Pass exact after rehearsal: 2026-10-06T16:58:20.012Z to 2026-10-06T16:59:20.013Z (60.0 s), 10,446 searches, 0 failed, p50 5.668 ms, mean 5.742 ms, p99 7.780 ms, 174.1 QPS (1000/QPS 5.744 ms).",
        "Pass default@1 after exact: 2026-10-06T16:59:38.026Z to 2026-10-06T16:59:58.028Z (20.0 s), 4,716 searches, 0 failed, p50 4.186 ms, mean 4.239 ms, p99 4.965 ms, 235.8 QPS (1000/QPS 4.241 ms).",
        "Pass default@8 after default@1: 2026-10-06T17:00:16.050Z to 2026-10-06T17:00:36.057Z (20.0 s), 11,527 searches, 0 failed, p50 13.343 ms, mean 13.882 ms, p99 25.976 ms, 576.1 QPS (1000/QPS 1.736 ms).",
        "Passes in the order run: exact, default@1, default@8; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 47,273 sent, 0 failed (warmupErrors). Index after the searches: ready, 524 of 524 indexed (HNSW graph returns 524 of 524 vectors, no writes queued: num_documents 524, pending_write_batches 0, health ok, unfiltered vector query with k above the count returned 524; the graph is built in memory during each write, so there is no later build step).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-typesense (8ca3b00f385a) on network engines_default at container-ip:8108, not through docker-proxy localhost:8108. Open after the passes: container-ip:8108 x8 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-typesense: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-typesense: 303 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): exact 3492/3492 MHz, average 3492.7/3492.9, performance, outside load 0.14 before the warm-up, 0.21 during, client CPU 0.540 ms per search; default@1 3492/3492 MHz, average 3492.4/3492.8, performance, outside load 0.21 before the warm-up, 0.20 during, client CPU 0.521 ms per search; default@8 3492/3492 MHz, average 3492.0/3492.5, performance, outside load 0.21 before the warm-up, 0.05 during, client CPU 0.592 ms per search."
      ]
    },
    {
      "name": "mongodb",
      "engine": "MongoDB 8.0.32 Atlas Local (mongod \u002B mongot Vector Search)",
      "index": "vectorSearch index, HNSW maxEdges=16 numEdgeCandidates=128, float32 binData, cosine, numCandidates=20x hits (min 100); exact mode = $vectorSearch exact:true",
      "hosting": "compose",
      "searchSettings": {
        "maxEdges": "16",
        "numCandidates": "20x",
        "numEdgeCandidates": "128"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.1684566,
        "rowsPerSecond": 3110.593470365661,
        "countMatchSeconds": 2.2935671,
        "indexSeconds": 10.234526,
        "indexNote": "index status READY, queryable True; mongot holds 524 of 524 documents in 4 segment(s) (196 \u002B 175 \u002B 139 \u002B 14 documents), 4 searched through the HNSW graph (Approximate); segment layout unchanged for 10.2 s over 11 readings; wait took 10.2 s"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 15409,
        "p50Ms": 1.262,
        "p95Ms": 1.4899799999999996,
        "p99Ms": 1.847368,
        "qps": {
          "1": 770.4190522666705,
          "8": 2351.875618430976
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 47194,
        "exactP50Ms": 1.2424,
        "exactP95Ms": 1.4585,
        "exactRecall": 1,
        "clientCpuMsPerSearch": {
          "1": 0.3031,
          "8": 0.2912
        }
      },
      "ram": {
        "bytes": 668676915,
        "text": "637.7 MiB (docker stats)"
      },
      "disk": {
        "bytes": 4812,
        "text": "4.7 KiB added by this load (1.37 GiB (whole engine data folder), 1.37 GiB before)"
      },
      "passOrder": [
        "default@1",
        "default@8",
        "exact"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "index status READY, queryable True; mongot holds 524 of 524 documents in 4 segment(s) (196 \u002B 175 \u002B 139 \u002B 14 documents), 4 searched through the HNSW graph (Approximate)"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "index status READY, queryable True; mongot holds 524 of 524 documents in 4 segment(s) (196 \u002B 175 \u002B 139 \u002B 14 documents), 4 searched through the HNSW graph (Approximate)"
        }
      },
      "durability": "Acknowledged writes are journaled before the acknowledgement. The sink sets no write concern, so the server default applies: getDefaultRWConcern gives w majority and the one-member replica set (--replSet gvbmongo in the image) has writeConcernMajorityJournalDefault true, with WiredTiger journaling on (journalCommitInterval 100 ms is only the interval for unacknowledged work). Measured 2026-10-04: 30 acknowledged single-document inserts raised WiredTiger \u0027log sync operations\u0027 by 30 and strace showed fdatasync on /data/db/journal/WiredTigerLog files. A crash loses no acknowledged write. mongot\u0027s search index is not part of that promise: it follows the collection asynchronously and is rebuilt from it.",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started mongodb.compose.yaml with every container created on CPUs 2-3,6-7: gvb-mongodb cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [mongodb.compose.yaml 4ee6f6411f1f]. This run created the engine from these files.",
        "Index after the load: ready, 524 of 524 indexed (index status READY, queryable True; mongot holds 524 of 524 documents in 4 segment(s) (196 \u002B 175 \u002B 139 \u002B 14 documents), 4 searched through the HNSW graph (Approximate)). Durability: Acknowledged writes are journaled before the acknowledgement. The sink sets no write concern, so the server default applies: getDefaultRWConcern gives w majority and the one-member replica set (--replSet gvbmongo in the image) has writeConcernMajorityJournalDefault true, with WiredTiger journaling on (journalCommitInterval 100 ms is only the interval for unacknowledged work). Measured 2026-10-04: 30 acknowledged single-document inserts raised WiredTiger \u0027log sync operations\u0027 by 30 and strace showed fdatasync on /data/db/journal/WiredTigerLog files. A crash loses no acknowledged write. mongot\u0027s search index is not part of that promise: it follows the collection asynchronously and is rebuilt from it..",
        "Outside load when searching began: 0.22 CPUs on average over the last 14 s and 0.20 over the last few seconds (limit 0.3), a quiet box. The load average 3.29 2.79 3.03 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@1 19,579 searches in 30.0 s, default@8 66,840 searches in 30.0 s, exact 23,155 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@1: warm-up 15.0 s and 11,583 searches (0 failed) at 1 searcher; p50 of the last windows 1.246, 1.267, 1.256 ms (windows of at least 2 s and 100 searches); trial of 2,307 searches in 3.0 s at 1 searcher: p50 1.265 ms against the settled p50 1.256 ms, 1% apart (limit 10%); the timed pass p50 1.262 ms, 0% from the settled p50 1.256 ms (limit 10% or 0.1 ms; 0.006 ms); default@8: warm-up 15.0 s and 34,974 searches (0 failed) at 8 searchers; QPS of the last windows 2,368, 2,359, 2,346 (windows of at least 2 s and 100 searches); trial of 7,033 searches in 3.0 s at 8 searchers: 2,342 QPS against the settled 2,359 QPS, 1% apart (limit 10%); the timed pass 2,352 QPS, 0% from the settled 2,359 QPS (limit 10%); exact: warm-up 15.0 s and 11,790 searches (0 failed) at 1 searcher; p50 of the last windows 1.246, 1.244, 1.251 ms (windows of at least 2 s and 100 searches); trial of 2,372 searches in 3.0 s at 1 searcher: p50 1.235 ms against the settled p50 1.246 ms, 1% apart (limit 10%); the timed pass p50 1.242 ms, 0% from the settled p50 1.246 ms (limit 10% or 0.1 ms; 0.004 ms).",
        "Pass default@1 after rehearsal: 2026-10-06T17:02:47.896Z to 2026-10-06T17:03:07.897Z (20.0 s), 15,409 searches, 0 failed, p50 1.262 ms, mean 1.296 ms, p99 1.847 ms, 770.4 QPS (1000/QPS 1.298 ms).",
        "Pass default@8 after default@1: 2026-10-06T17:03:36.920Z to 2026-10-06T17:03:56.922Z (20.0 s), 47,043 searches, 0 failed, p50 3.282 ms, mean 3.399 ms, p99 6.078 ms, 2351.9 QPS (1000/QPS 0.425 ms).",
        "Pass exact after default@8: 2026-10-06T17:04:14.939Z to 2026-10-06T17:05:14.940Z (60.0 s), 47,194 searches, 0 failed, p50 1.242 ms, mean 1.269 ms, p99 1.791 ms, 786.6 QPS (1000/QPS 1.271 ms).",
        "Passes in the order run: default@1, default@8, exact; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 179,633 sent, 0 failed (warmupErrors). Index after the searches: ready, 524 of 524 indexed (index status READY, queryable True; mongot holds 524 of 524 documents in 4 segment(s) (196 \u002B 175 \u002B 139 \u002B 14 documents), 4 searched through the HNSW graph (Approximate)).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-mongodb (30db26986ada) on network gvb-mongodb_default at container-ip:27017, not through docker-proxy localhost:27017. Open after the passes: container-ip:27017 x10 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-mongodb: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-mongodb: 210 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@1 3492/3492 MHz, average 3492.3/3492.4, performance, outside load 0.27 before the warm-up after waiting 1 s, 0.35 during, BUSY BOX, client CPU 0.303 ms per search; default@8 3492/3492 MHz, average 3492.0/3492.2, performance, outside load 0.30 before the warm-up after waiting 11 s, 0.25 during, client CPU 0.291 ms per search; exact 3492/3492 MHz, average 3492.0/3492.2, performance, outside load 0.24 before the warm-up, 0.30 during, BUSY BOX, client CPU 0.302 ms per search.",
        "WARNING: busy box during default@1: processes outside the benchmark used 0.35 CPUs on average during the pass, above the limit of 0.3.",
        "WARNING: busy box during exact: processes outside the benchmark used 0.30 CPUs on average during the pass, above the limit of 0.3."
      ]
    },
    {
      "name": "milvus",
      "engine": "Milvus 2.6.25 (standalone, embedded etcd, local storage, REST API v2)",
      "index": "HNSW M=16 efConstruction=128, ef=100, metric COSINE, Strong consistency searches; approximate only (no exact mode)",
      "hosting": "compose",
      "searchSettings": {
        "M": "16",
        "ef": "100",
        "efConstruction": "128"
      },
      "load": {
        "rows": 524,
        "batch": 1000,
        "upsertSeconds": 0.4475171,
        "rowsPerSecond": 1170.9049777092316,
        "countMatchSeconds": 1.6960359,
        "indexSeconds": 138.5631972,
        "indexNote": "index Finished, type HNSW (COSINE, {\u0022M\u0022:16,\u0022efConstruction\u0022:128}), indexedRows 524 of 524 sealed rows, pendingRows 0; stored rows 524; LoadStateLoaded; query node: 1 sealed segment(s) with the index loaded covering 524 rows against 524 stored rows, 0 sealed without it, 0 rows in growing segments; datacoord: 1 live segment(s), 1 flushed and indexed covering 524 rows, 0 delete-log (L0) segment(s) waiting for a compaction, 3 segment(s) compacted away so far; flush, index build and load took 14.6 s, then compaction: 1 job(s) accepted by Milvus, layout ready and unchanged for 75 s, 123.9 s for the whole step; search ledger started, later state reads count the searches against the query node\u0027s own counters"
      },
      "search": {
        "countInTarget": 524,
        "latencySamples": 8145,
        "p50Ms": 2.3231,
        "p95Ms": 2.90616,
        "p99Ms": 3.712779999999996,
        "qps": {
          "1": 407.24678682285196,
          "8": 1262.7879149446694
        },
        "recall": 1,
        "ndcg": 0.5181699774768911,
        "errors": 0,
        "exactQueries": 0,
        "clientCpuMsPerSearch": {
          "1": 0.5687,
          "8": 0.6288
        }
      },
      "ram": {
        "bytes": 230791577,
        "text": "220.1 MiB (docker stats)"
      },
      "disk": {
        "bytes": 11400822,
        "text": "10.87 MiB added by this load (220 MiB (whole engine data folder), 209.13 MiB before)"
      },
      "passOrder": [
        "default@1",
        "default@8"
      ],
      "warmupErrors": 0,
      "indexState": {
        "afterLoad": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "index Finished, type HNSW (COSINE, {\u0022M\u0022:16,\u0022efConstruction\u0022:128}), indexedRows 524 of 524 sealed rows, pendingRows 0; stored rows 524; LoadStateLoaded; query node: 1 sealed segment(s) with the index loaded covering 524 rows against 524 stored rows, 0 sealed without it, 0 rows in growing segments; datacoord: 1 live segment(s), 1 flushed and indexed covering 524 rows, 0 delete-log (L0) segment(s) waiting for a compaction, 3 segment(s) compacted away so far; search ledger since the index finished: the sink sent 0 searches (searchParams.params.ef=100, Strong consistency), the query node counted 0 search requests for this collection, segment searches Sealed \u002B0 and Growing \u002B0, the same sealed segments with the same loaded index builds at both readings, segments compacted away \u002B0"
        },
        "afterSearch": {
          "ready": true,
          "indexedVectors": 524,
          "totalVectors": 524,
          "detail": "index Finished, type HNSW (COSINE, {\u0022M\u0022:16,\u0022efConstruction\u0022:128}), indexedRows 524 of 524 sealed rows, pendingRows 0; stored rows 524; LoadStateLoaded; query node: 1 sealed segment(s) with the index loaded covering 524 rows against 524 stored rows, 0 sealed without it, 0 rows in growing segments; datacoord: 1 live segment(s), 1 flushed and indexed covering 524 rows, 0 delete-log (L0) segment(s) waiting for a compaction, 3 segment(s) compacted away so far; search ledger since the index finished: the sink sent 113366 searches (searchParams.params.ef=100, Strong consistency), the query node counted 113366 search requests for this collection, segment searches Sealed \u002B111237 and Growing \u002B0, the same sealed segments with the same loaded index builds at both readings, segments compacted away \u002B0"
        }
      },
      "durability": "Writes are not fsynced. Standalone uses the default message queue rocksmq (RocksDB, /var/lib/milvus/rdb_data, mq.type default) and flushed segments go to local-disk object storage (COMMON_STORAGETYPE=local in milvus.compose.yaml). Measured 2026-10-04 with strace over 30 acknowledged single-row upserts and one flush: fdatasync ran only on the embedded etcd files (member/wal, member/snap/db), never on rdb_data or the segment files. A host power loss or kernel crash can lose acknowledged rows that are still in the OS page cache; a Milvus process crash alone should not (inferred, not tested). Metadata in embedded etcd is fdatasynced on every commit.",
      "notes": [
        "Started by run-all for this measurement with every container created on CPUs 2-3,6-7 (started milvus.compose.yaml with every container created on CPUs 2-3,6-7: gvb-milvus cpuset 2-3,6-7 (read back from docker inspect)); stopped afterwards.",
        "Engine files (SHA-256, first 12 hex digits): [milvus.compose.yaml eaefecdc3671]. This run created the engine from these files.",
        "Index after the load: ready, 524 of 524 indexed (index Finished, type HNSW (COSINE, {\u0022M\u0022:16,\u0022efConstruction\u0022:128}), indexedRows 524 of 524 sealed rows, pendingRows 0; stored rows 524; LoadStateLoaded; query node: 1 sealed segment(s) with the index loaded covering 524 rows against 524 stored rows, 0 sealed without it, 0 rows in growing segments; datacoord: 1 live segment(s), 1 flushed and indexed covering 524 rows, 0 delete-log (L0) segment(s) waiting for a compaction, 3 segment(s) compacted away so far; search ledger since the index finished: the sink sent 0 searches (searchParams.params.ef=100, Strong consistency), the query node counted 0 search requests for this collection, segment searches Sealed \u002B0 and Growing \u002B0, the same sealed segments with the same loaded index builds at both readings, segments compacted away \u002B0). Durability: Writes are not fsynced. Standalone uses the default message queue rocksmq (RocksDB, /var/lib/milvus/rdb_data, mq.type default) and flushed segments go to local-disk object storage (COMMON_STORAGETYPE=local in milvus.compose.yaml). Measured 2026-10-04 with strace over 30 acknowledged single-row upserts and one flush: fdatasync ran only on the embedded etcd files (member/wal, member/snap/db), never on rdb_data or the segment files. A host power loss or kernel crash can lose acknowledged rows that are still in the OS page cache; a Milvus process crash alone should not (inferred, not tested). Metadata in embedded etcd is fdatasynced on every commit..",
        "Outside load when searching began: 0.19 CPUs on average over the last 60 s and 0.20 over the last few seconds (limit 0.3), a quiet box. The load average 0.30 1.57 2.49 (1/5/15 min; it also counts this benchmark\u0027s own client and engines, so it is recorded, not judged).",
        "Rehearsal before any timed pass, untimed, every pass type at its own concurrency for at least 30 s through the same code the timed passes use: default@1 12,177 searches in 30.0 s, default@8 37,782 searches in 30.0 s.",
        "Settle: settled before every timed pass, each by its own warm-up at its own concurrency and a trial of the same pass within 10% of the warm-up\u0027s settled figure, and the timed pass itself within 10% of it. default@1: warm-up 15.0 s and 6,114 searches (0 failed) at 1 searcher; p50 of the last windows 2.323, 2.323, 2.319 ms (windows of at least 2 s and 100 searches); trial of 1,228 searches in 3.0 s at 1 searcher: p50 2.316 ms against the settled p50 2.323 ms, 0% apart (limit 10%); the timed pass p50 2.323 ms, 0% from the settled p50 2.323 ms (limit 10% or 0.1 ms; 0.000 ms); default@8: warm-up 15.0 s and 18,865 searches (0 failed) at 8 searchers; QPS of the last windows 1,247, 1,268, 1,271 (windows of at least 2 s and 100 searches); trial of 3,797 searches in 3.0 s at 8 searchers: 1,264 QPS against the settled 1,268 QPS, 0% apart (limit 10%); the timed pass 1,263 QPS, 0% from the settled 1,268 QPS (limit 10%).",
        "Pass default@1 after rehearsal: 2026-10-06T17:10:03.627Z to 2026-10-06T17:10:23.627Z (20.0 s), 8,145 searches, 0 failed, p50 2.323 ms, mean 2.453 ms, p99 3.713 ms, 407.2 QPS (1000/QPS 2.456 ms).",
        "Pass default@8 after default@1: 2026-10-06T17:10:41.641Z to 2026-10-06T17:11:01.643Z (20.0 s), 25,258 searches, 0 failed, p50 5.795 ms, mean 6.332 ms, p99 14.322 ms, 1262.8 QPS (1000/QPS 0.792 ms).",
        "Passes in the order run: default@1, default@8; each after its own warm-up. Untimed searches (rehearsals, per-pass warm-ups, settle checks and extensions): 79,963 sent, 0 failed (warmupErrors). Index after the searches: ready, 524 of 524 indexed (index Finished, type HNSW (COSINE, {\u0022M\u0022:16,\u0022efConstruction\u0022:128}), indexedRows 524 of 524 sealed rows, pendingRows 0; stored rows 524; LoadStateLoaded; query node: 1 sealed segment(s) with the index loaded covering 524 rows against 524 stored rows, 0 sealed without it, 0 rows in growing segments; datacoord: 1 live segment(s), 1 flushed and indexed covering 524 rows, 0 delete-log (L0) segment(s) waiting for a compaction, 3 segment(s) compacted away so far; search ledger since the index finished: the sink sent 113366 searches (searchParams.params.ef=100, Strong consistency), the query node counted 113366 search requests for this collection, segment searches Sealed \u002B111237 and Growing \u002B0, the same sealed segments with the same loaded index builds at both readings, segments compacted away \u002B0).",
        "Benchmark copy gvbbench_eshoponweb dropped afterwards.",
        "Connection: container address on its Docker network, not the published port (no docker-proxy): container gvb-milvus (6a4ba4e1d723) on network gvb-milvus_default at container-ip:19530, not through docker-proxy localhost:19530; container gvb-milvus (6a4ba4e1d723) on network gvb-milvus_default at container-ip:9091, not through docker-proxy localhost:9091. Open after the passes: container-ip:19530 x8 (this target).",
        "CPU pinning: cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which), CPUs 2-3,6-7. Changed: container gvb-milvus: already on CPUs 2-3,6-7 (its cpuset since creation), not changed. Read back: container gvb-milvus: 24 thread(s) on CPUs 2-3,6-7 (every thread of the container).",
        "Clock per pass (median MHz of engine CPUs / client CPUs, their average MHz, governor, CPUs busy outside the benchmark, client CPU per search; the clock was pinned at 3500 MHz (turbo off), and a pass whose average on either group is more than 1% off it is flagged): default@1 3492/3492 MHz, average 3492.0/3492.3, performance, outside load 0.22 before the warm-up, 0.20 during, client CPU 0.569 ms per search; default@8 3492/3492 MHz, average 3492.0/3492.3, performance, outside load 0.21 before the warm-up, 0.25 during, client CPU 0.629 ms per search."
      ]
    }
  ],
  "conditions": {
    "buildConfiguration": "Release",
    "jitOptimizerDisabled": false,
    "dotNet": ".NET 10.0.12",
    "runtimeVersion": "10.0.12",
    "serverGc": false,
    "machineControl": "on",
    "governor": "performance",
    "governorBefore": "schedutil",
    "governorAtEnd": "performance",
    "governorAfterRestore": "schedutil",
    "logicalCpus": 8,
    "threadSiblings": [
      "0,4",
      "1,5",
      "2,6",
      "3,7"
    ],
    "clientCpus": "0-1,4-5",
    "engineCpus": "2-3,6-7",
    "clientPinning": "client pid 319433: 22 thread(s) on CPUs 0-1,4-5",
    "warmupSearches": 20,
    "exactSeconds": 60,
    "search": {
      "top": 10,
      "queries": "golden",
      "concurrency": [
        1,
        8
      ],
      "secondsPerLevel": 20,
      "warmupSearches": 20,
      "exactSeconds": 60,
      "searchTimeoutSeconds": 120,
      "batch": 1000,
      "seed": 703
    },
    "busyRule": "right before each pass\u0027s warm-up the run waits, up to 10 min, while processes outside the benchmark (everything but this client and the engine under test\u0027s cgroups) use more than 0.3 CPUs on average over the last 60 s or the last 5 s, neither window reaching back past the start of the target or of its engine (so the engine\u0027s own start-up is not outside work); if the box does not clear the pass runs anyway, flagged \u0027busy box\u0027, as is a pass whose own outside load is above the limit",
    "engines": [
      {
        "target": "sqlitevec",
        "hosting": "embedded",
        "method": "none: an embedded engine runs inside the benchmark client process, so it shares the client CPUs",
        "cpus": "0-1,4-5",
        "changes": [],
        "verified": [],
        "problems": [],
        "restored": []
      },
      {
        "target": "mariadb",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvbbench-mariadb: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvbbench-mariadb: 13 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of mariadb-bench.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "duckdb",
        "hosting": "embedded",
        "method": "none: an embedded engine runs inside the benchmark client process, so it shares the client CPUs",
        "cpus": "0-1,4-5",
        "changes": [],
        "verified": [],
        "problems": [],
        "restored": []
      },
      {
        "target": "qdrant",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-qdrant: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-qdrant: 24 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of qdrant.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "elasticsearch",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-elasticsearch: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-elasticsearch: 75 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of elasticsearch.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "vespa",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-vespa: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-vespa: 214 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of vespa.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "qdrant-hnsw",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-qdrant: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-qdrant: 24 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of qdrant.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "sql-diskann",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-mssql: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-mssql: 84 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of mssql.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "pgvector",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-pgvector: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-pgvector: 6 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of pgvector.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "weaviate",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-weaviate: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-weaviate: 9 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of weaviate.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "clickhouse",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-clickhouse: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-clickhouse: 723 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of clickhouse.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "oracle",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-oracle: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-oracle: 85 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of oracle.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "sql",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-mssql: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-mssql: 86 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of mssql.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "opensearch",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-opensearch: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-opensearch: 55 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of opensearch.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "redis",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-redis: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-redis: 16 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of redis.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "chroma",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-chroma: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-chroma: 9 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of chroma.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "typesense",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-typesense: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-typesense: 303 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of typesense.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "mongodb",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-mongodb: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-mongodb: 210 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of mongodb.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      },
      {
        "target": "milvus",
        "hosting": "compose",
        "method": "cpuset from the container\u0027s creation when run-all started it on the engine CPUs, else docker update --cpuset-cpus (each container\u0027s change line says which)",
        "cpus": "2-3,6-7",
        "changes": [
          "container gvb-milvus: already on CPUs 2-3,6-7 (its cpuset since creation), not changed"
        ],
        "verified": [
          "container gvb-milvus: 24 thread(s) on CPUs 2-3,6-7 (every thread of the container)"
        ],
        "problems": [],
        "restored": [
          "no container of milvus.compose.yaml is left on CPUs 2-3,6-7; nothing to put back"
        ]
      }
    ],
    "passes": [
      {
        "target": "sqlitevec",
        "pass": "default@1",
        "startUtc": "2026-10-06T15:50:27.518Z",
        "endUtc": "2026-10-06T15:50:47.567Z",
        "seconds": 20.049,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 81,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3476,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 1,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3486,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3492,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3474,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T15:50:09.489Z",
        "quietCheckLeadSeconds": 18.029,
        "outsideWindowFromUtc": "2026-10-06T15:49:09.489Z",
        "outsideLoadAtStart": 0.246,
        "outsideLoadRecentAtStart": 0.219,
        "outsideLoadDuring": 0.212,
        "accountingMismatch": false,
        "searches": 10129,
        "clientCpuSeconds": 20.9314,
        "clientCpuMsPerSearch": 2.0665,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T15:50:27.519Z to 2026-10-06T15:50:47.519Z, 80 samples, divided by the pass\u0027s 10129 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "sqlitevec",
        "pass": "default@8",
        "startUtc": "2026-10-06T15:51:05.588Z",
        "endUtc": "2026-10-06T15:51:25.597Z",
        "seconds": 20.009,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 76,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 1,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 2,
            "min": 3491,
            "median": 3492,
            "max": 3542
          },
          {
            "cpu": 3,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 5,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 6,
            "min": 3490,
            "median": 3492,
            "max": 3528
          },
          {
            "cpu": 7,
            "min": 3453,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T15:50:47.580Z",
        "quietCheckLeadSeconds": 18.007,
        "outsideWindowFromUtc": "2026-10-06T15:49:47.580Z",
        "outsideLoadAtStart": 0.243,
        "outsideLoadRecentAtStart": 0.194,
        "outsideLoadDuring": 0.229,
        "accountingMismatch": false,
        "searches": 22771,
        "clientCpuSeconds": 79.7374,
        "clientCpuMsPerSearch": 3.5017,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T15:51:05.588Z to 2026-10-06T15:51:25.591Z, 76 samples, divided by the pass\u0027s 22771 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "sqlitevec",
        "pass": "exact",
        "startUtc": "2026-10-06T15:51:43.600Z",
        "endUtc": "2026-10-06T15:52:43.629Z",
        "seconds": 60.029,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 1,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3400,
            "median": 3492,
            "max": 3603
          },
          {
            "cpu": 3,
            "min": 3450,
            "median": 3492,
            "max": 3563
          },
          {
            "cpu": 4,
            "min": 3465,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3420,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3487,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3487,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T15:51:25.597Z",
        "quietCheckLeadSeconds": 18.003,
        "outsideWindowFromUtc": "2026-10-06T15:50:25.597Z",
        "outsideLoadAtStart": 0.229,
        "outsideLoadRecentAtStart": 0.216,
        "outsideLoadDuring": 0.221,
        "accountingMismatch": false,
        "searches": 30398,
        "clientCpuSeconds": 62.8781,
        "clientCpuMsPerSearch": 2.0685,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T15:51:43.600Z to 2026-10-06T15:52:43.602Z, 240 samples, divided by the pass\u0027s 30398 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "mariadb",
        "pass": "default@1",
        "startUtc": "2026-10-06T15:54:39.440Z",
        "endUtc": "2026-10-06T15:54:59.472Z",
        "seconds": 20.032,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3488,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 1,
            "min": 3490,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 2,
            "min": 3490,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3491,
            "median": 3492,
            "max": 3506
          },
          {
            "cpu": 4,
            "min": 3490,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3488,
            "median": 3492,
            "max": 3503
          },
          {
            "cpu": 7,
            "min": 3488,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T15:54:21.434Z",
        "quietCheckLeadSeconds": 18.005,
        "outsideWindowFromUtc": "2026-10-06T15:53:21.434Z",
        "outsideLoadAtStart": 0.178,
        "outsideLoadRecentAtStart": 0.165,
        "outsideLoadDuring": 0.201,
        "accountingMismatch": false,
        "searches": 31037,
        "clientCpuSeconds": 14.9519,
        "clientCpuMsPerSearch": 0.4817,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T15:54:39.440Z to 2026-10-06T15:54:59.440Z, 80 samples, divided by the pass\u0027s 31037 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "mariadb",
        "pass": "default@8",
        "startUtc": "2026-10-06T15:55:17.476Z",
        "endUtc": "2026-10-06T15:55:37.518Z",
        "seconds": 20.042,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3492,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 1,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 5,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T15:54:59.472Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T15:53:59.472Z",
        "outsideLoadAtStart": 0.19,
        "outsideLoadRecentAtStart": 0.219,
        "outsideLoadDuring": 0.17,
        "accountingMismatch": false,
        "searches": 118362,
        "clientCpuSeconds": 44.2262,
        "clientCpuMsPerSearch": 0.3737,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T15:55:17.476Z to 2026-10-06T15:55:37.477Z, 80 samples, divided by the pass\u0027s 118362 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "mariadb",
        "pass": "exact",
        "startUtc": "2026-10-06T15:55:55.523Z",
        "endUtc": "2026-10-06T15:56:55.560Z",
        "seconds": 60.037,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 2177,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 1,
            "min": 2195,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3465,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 6,
            "min": 3488,
            "median": 3492,
            "max": 3503
          },
          {
            "cpu": 7,
            "min": 3489,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T15:55:37.519Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T15:54:37.519Z",
        "outsideLoadAtStart": 0.182,
        "outsideLoadRecentAtStart": 0.153,
        "outsideLoadDuring": 0.197,
        "accountingMismatch": false,
        "searches": 35899,
        "clientCpuSeconds": 17.8572,
        "clientCpuMsPerSearch": 0.4974,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T15:55:55.523Z to 2026-10-06T15:56:55.523Z, 240 samples, divided by the pass\u0027s 35899 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "duckdb",
        "pass": "default@8",
        "startUtc": "2026-10-06T15:58:45.775Z",
        "endUtc": "2026-10-06T15:59:05.786Z",
        "seconds": 20.011,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 77,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 1,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 2,
            "min": 3490,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 3,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 5,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 6,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3491,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T15:58:27.754Z",
        "quietCheckLeadSeconds": 18.02,
        "outsideWindowFromUtc": "2026-10-06T15:57:27.754Z",
        "outsideLoadAtStart": 0.206,
        "outsideLoadRecentAtStart": 0.196,
        "outsideLoadDuring": 0.225,
        "accountingMismatch": false,
        "searches": 11525,
        "clientCpuSeconds": 79.5488,
        "clientCpuMsPerSearch": 6.9023,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T15:58:45.775Z to 2026-10-06T15:59:05.783Z, 77 samples, divided by the pass\u0027s 11525 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "duckdb",
        "pass": "exact",
        "startUtc": "2026-10-06T15:59:23.792Z",
        "endUtc": "2026-10-06T16:00:23.807Z",
        "seconds": 60.014,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 1,
            "min": 3489,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 2,
            "min": 3463,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3485,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3505
          },
          {
            "cpu": 6,
            "min": 3486,
            "median": 3492,
            "max": 3549
          },
          {
            "cpu": 7,
            "min": 3490,
            "median": 3492,
            "max": 3548
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T15:59:05.786Z",
        "quietCheckLeadSeconds": 18.006,
        "outsideWindowFromUtc": "2026-10-06T15:58:05.786Z",
        "outsideLoadAtStart": 0.205,
        "outsideLoadRecentAtStart": 0.237,
        "outsideLoadDuring": 0.214,
        "accountingMismatch": false,
        "searches": 14028,
        "clientCpuSeconds": 87.6864,
        "clientCpuMsPerSearch": 6.2508,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T15:59:23.792Z to 2026-10-06T16:00:23.796Z, 240 samples, divided by the pass\u0027s 14028 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "duckdb",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:00:41.813Z",
        "endUtc": "2026-10-06T16:01:01.821Z",
        "seconds": 20.009,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 1,
            "min": 3492,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 2,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3490,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 6,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:00:23.807Z",
        "quietCheckLeadSeconds": 18.005,
        "outsideWindowFromUtc": "2026-10-06T15:59:23.807Z",
        "outsideLoadAtStart": 0.214,
        "outsideLoadRecentAtStart": 0.221,
        "outsideLoadDuring": 0.185,
        "accountingMismatch": false,
        "searches": 5404,
        "clientCpuSeconds": 27.1105,
        "clientCpuMsPerSearch": 5.0167,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:00:41.813Z to 2026-10-06T16:01:01.816Z, 80 samples, divided by the pass\u0027s 5404 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "qdrant",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:02:27.715Z",
        "endUtc": "2026-10-06T16:02:47.739Z",
        "seconds": 20.023,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3491,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 1,
            "min": 3492,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3492,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:02:09.711Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T16:01:09.711Z",
        "outsideLoadAtStart": 0.211,
        "outsideLoadRecentAtStart": 0.149,
        "outsideLoadDuring": 0.197,
        "accountingMismatch": false,
        "searches": 65748,
        "clientCpuSeconds": 34.4196,
        "clientCpuMsPerSearch": 0.5235,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:02:27.715Z to 2026-10-06T16:02:47.718Z, 80 samples, divided by the pass\u0027s 65748 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "qdrant",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:03:05.742Z",
        "endUtc": "2026-10-06T16:03:25.766Z",
        "seconds": 20.024,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3491,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 1,
            "min": 3474,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 4,
            "min": 3492,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 5,
            "min": 3490,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:02:47.739Z",
        "quietCheckLeadSeconds": 18.003,
        "outsideWindowFromUtc": "2026-10-06T16:01:47.739Z",
        "outsideLoadAtStart": 0.203,
        "outsideLoadRecentAtStart": 0.139,
        "outsideLoadDuring": 0.17,
        "accountingMismatch": false,
        "searches": 22368,
        "clientCpuSeconds": 16.1665,
        "clientCpuMsPerSearch": 0.7228,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:03:05.742Z to 2026-10-06T16:03:25.743Z, 80 samples, divided by the pass\u0027s 22368 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "elasticsearch",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:05:48.819Z",
        "endUtc": "2026-10-06T16:06:08.837Z",
        "seconds": 20.018,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3491,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 1,
            "min": 3491,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3492,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 5,
            "min": 3490,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:05:30.816Z",
        "quietCheckLeadSeconds": 18.003,
        "outsideWindowFromUtc": "2026-10-06T16:04:30.816Z",
        "outsideLoadAtStart": 0.209,
        "outsideLoadRecentAtStart": 0.205,
        "outsideLoadDuring": 0.251,
        "accountingMismatch": false,
        "searches": 54163,
        "clientCpuSeconds": 26.5892,
        "clientCpuMsPerSearch": 0.4909,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:05:48.819Z to 2026-10-06T16:06:08.821Z, 80 samples, divided by the pass\u0027s 54163 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "elasticsearch",
        "pass": "exact",
        "startUtc": "2026-10-06T16:06:26.840Z",
        "endUtc": "2026-10-06T16:07:26.884Z",
        "seconds": 60.044,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3489,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 1,
            "min": 3468,
            "median": 3492,
            "max": 3524
          },
          {
            "cpu": 2,
            "min": 3489,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 3,
            "min": 3475,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3487,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3425,
            "median": 3492,
            "max": 3502
          },
          {
            "cpu": 6,
            "min": 3424,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 7,
            "min": 3372,
            "median": 3492,
            "max": 3499
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:06:08.838Z",
        "quietCheckLeadSeconds": 18.003,
        "outsideWindowFromUtc": "2026-10-06T16:05:08.838Z",
        "outsideLoadAtStart": 0.229,
        "outsideLoadRecentAtStart": 0.211,
        "outsideLoadDuring": 0.233,
        "accountingMismatch": false,
        "searches": 46677,
        "clientCpuSeconds": 21.5176,
        "clientCpuMsPerSearch": 0.461,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:06:26.840Z to 2026-10-06T16:07:26.841Z, 240 samples, divided by the pass\u0027s 46677 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "elasticsearch",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:07:44.888Z",
        "endUtc": "2026-10-06T16:08:04.903Z",
        "seconds": 20.015,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 1,
            "min": 3438,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3426,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 4,
            "min": 3473,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3388,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3487,
            "median": 3492,
            "max": 3493
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:07:26.885Z",
        "quietCheckLeadSeconds": 18.003,
        "outsideWindowFromUtc": "2026-10-06T16:06:26.885Z",
        "outsideLoadAtStart": 0.233,
        "outsideLoadRecentAtStart": 0.212,
        "outsideLoadDuring": 0.216,
        "accountingMismatch": false,
        "searches": 14962,
        "clientCpuSeconds": 6.8591,
        "clientCpuMsPerSearch": 0.4584,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:07:44.888Z to 2026-10-06T16:08:04.889Z, 80 samples, divided by the pass\u0027s 14962 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "vespa",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:10:52.850Z",
        "endUtc": "2026-10-06T16:11:12.859Z",
        "seconds": 20.009,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3491,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 1,
            "min": 3491,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3488,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:10:34.848Z",
        "quietCheckLeadSeconds": 18.002,
        "outsideWindowFromUtc": "2026-10-06T16:09:34.848Z",
        "outsideLoadAtStart": 0.226,
        "outsideLoadRecentAtStart": 0.213,
        "outsideLoadDuring": 0.202,
        "accountingMismatch": false,
        "searches": 9919,
        "clientCpuSeconds": 9.2966,
        "clientCpuMsPerSearch": 0.9373,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:10:52.850Z to 2026-10-06T16:11:12.851Z, 80 samples, divided by the pass\u0027s 9919 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "vespa",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:11:30.866Z",
        "endUtc": "2026-10-06T16:11:50.877Z",
        "seconds": 20.011,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3491,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 1,
            "min": 3492,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3488,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:11:12.859Z",
        "quietCheckLeadSeconds": 18.007,
        "outsideWindowFromUtc": "2026-10-06T16:10:12.859Z",
        "outsideLoadAtStart": 0.192,
        "outsideLoadRecentAtStart": 0.199,
        "outsideLoadDuring": 0.258,
        "accountingMismatch": false,
        "searches": 31299,
        "clientCpuSeconds": 27.3983,
        "clientCpuMsPerSearch": 0.8754,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:11:30.866Z to 2026-10-06T16:11:50.869Z, 80 samples, divided by the pass\u0027s 31299 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "vespa",
        "pass": "exact",
        "startUtc": "2026-10-06T16:12:08.881Z",
        "endUtc": "2026-10-06T16:13:08.909Z",
        "seconds": 60.028,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 1,
            "min": 3489,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 4,
            "min": 3488,
            "median": 3492,
            "max": 3505
          },
          {
            "cpu": 5,
            "min": 3472,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:11:50.878Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T16:10:50.878Z",
        "outsideLoadAtStart": 0.241,
        "outsideLoadRecentAtStart": 0.261,
        "outsideLoadDuring": 0.184,
        "accountingMismatch": false,
        "searches": 29861,
        "clientCpuSeconds": 28.4133,
        "clientCpuMsPerSearch": 0.9515,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:12:08.881Z to 2026-10-06T16:13:08.883Z, 240 samples, divided by the pass\u0027s 29861 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "qdrant-hnsw",
        "pass": "exact",
        "startUtc": "2026-10-06T16:15:14.953Z",
        "endUtc": "2026-10-06T16:16:15.018Z",
        "seconds": 60.065,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3491,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 1,
            "min": 3490,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 2,
            "min": 3490,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 3,
            "min": 3452,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 4,
            "min": 3489,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 5,
            "min": 3427,
            "median": 3492,
            "max": 3502
          },
          {
            "cpu": 6,
            "min": 3491,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3493
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:14:56.950Z",
        "quietCheckLeadSeconds": 18.003,
        "outsideWindowFromUtc": "2026-10-06T16:13:56.950Z",
        "outsideLoadAtStart": 0.192,
        "outsideLoadRecentAtStart": 0.294,
        "outsideLoadDuring": 0.178,
        "accountingMismatch": false,
        "searches": 66375,
        "clientCpuSeconds": 48.1713,
        "clientCpuMsPerSearch": 0.7257,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:15:14.953Z to 2026-10-06T16:16:14.953Z, 240 samples, divided by the pass\u0027s 66375 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "qdrant-hnsw",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:16:33.022Z",
        "endUtc": "2026-10-06T16:16:53.042Z",
        "seconds": 20.02,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3437,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 1,
            "min": 3491,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 2,
            "min": 3491,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 3,
            "min": 3491,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 4,
            "min": 3459,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 6,
            "min": 3491,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3493
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:16:15.018Z",
        "quietCheckLeadSeconds": 18.003,
        "outsideWindowFromUtc": "2026-10-06T16:15:15.018Z",
        "outsideLoadAtStart": 0.178,
        "outsideLoadRecentAtStart": 0.164,
        "outsideLoadDuring": 0.16,
        "accountingMismatch": false,
        "searches": 21546,
        "clientCpuSeconds": 15.7319,
        "clientCpuMsPerSearch": 0.7302,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:16:33.022Z to 2026-10-06T16:16:53.022Z, 80 samples, divided by the pass\u0027s 21546 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "qdrant-hnsw",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:17:11.046Z",
        "endUtc": "2026-10-06T16:17:31.069Z",
        "seconds": 20.023,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3492,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 1,
            "min": 3492,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 5,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:16:53.042Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T16:15:53.042Z",
        "outsideLoadAtStart": 0.169,
        "outsideLoadRecentAtStart": 0.152,
        "outsideLoadDuring": 0.208,
        "accountingMismatch": false,
        "searches": 68972,
        "clientCpuSeconds": 35.2827,
        "clientCpuMsPerSearch": 0.5116,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:17:11.046Z to 2026-10-06T16:17:31.048Z, 80 samples, divided by the pass\u0027s 68972 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "sql-diskann",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:19:35.209Z",
        "endUtc": "2026-10-06T16:19:55.222Z",
        "seconds": 20.013,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 1,
            "min": 3489,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 2,
            "min": 3491,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 4,
            "min": 3487,
            "median": 3492,
            "max": 3502
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3503
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3494
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:19:17.185Z",
        "quietCheckLeadSeconds": 18.024,
        "outsideWindowFromUtc": "2026-10-06T16:18:17.185Z",
        "outsideLoadAtStart": 0.205,
        "outsideLoadRecentAtStart": 0.228,
        "outsideLoadDuring": 0.202,
        "accountingMismatch": false,
        "searches": 12914,
        "clientCpuSeconds": 13.9702,
        "clientCpuMsPerSearch": 1.0818,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:19:35.209Z to 2026-10-06T16:19:55.219Z, 80 samples, divided by the pass\u0027s 12914 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "sql-diskann",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:20:13.227Z",
        "endUtc": "2026-10-06T16:20:33.234Z",
        "seconds": 20.007,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3491,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 1,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3489,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 3,
            "min": 3487,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3487,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3492,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3489,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:19:55.223Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T16:18:55.223Z",
        "outsideLoadAtStart": 0.194,
        "outsideLoadRecentAtStart": 0.197,
        "outsideLoadDuring": 0.227,
        "accountingMismatch": false,
        "searches": 5359,
        "clientCpuSeconds": 5.4855,
        "clientCpuMsPerSearch": 1.0236,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:20:13.227Z to 2026-10-06T16:20:33.230Z, 80 samples, divided by the pass\u0027s 5359 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "sql-diskann",
        "pass": "exact",
        "startUtc": "2026-10-06T16:20:51.239Z",
        "endUtc": "2026-10-06T16:21:51.254Z",
        "seconds": 60.014,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3489,
            "median": 3492,
            "max": 3504
          },
          {
            "cpu": 1,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3483,
            "median": 3492,
            "max": 3502
          },
          {
            "cpu": 3,
            "min": 3477,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3461,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3429,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3488,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:20:33.235Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T16:19:33.235Z",
        "outsideLoadAtStart": 0.216,
        "outsideLoadRecentAtStart": 0.197,
        "outsideLoadDuring": 0.21,
        "accountingMismatch": false,
        "searches": 15442,
        "clientCpuSeconds": 17.8324,
        "clientCpuMsPerSearch": 1.1548,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:20:51.239Z to 2026-10-06T16:21:51.242Z, 240 samples, divided by the pass\u0027s 15442 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "pgvector",
        "pass": "exact",
        "startUtc": "2026-10-06T16:23:48.721Z",
        "endUtc": "2026-10-06T16:24:48.745Z",
        "seconds": 60.024,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3472,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 1,
            "min": 3474,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 2,
            "min": 3482,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3487,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3390,
            "median": 3492,
            "max": 3503
          },
          {
            "cpu": 7,
            "min": 3486,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:23:30.717Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T16:22:30.717Z",
        "outsideLoadAtStart": 0.209,
        "outsideLoadRecentAtStart": 0.196,
        "outsideLoadDuring": 0.201,
        "accountingMismatch": false,
        "searches": 26089,
        "clientCpuSeconds": 18.0217,
        "clientCpuMsPerSearch": 0.6908,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:23:48.721Z to 2026-10-06T16:24:48.723Z, 240 samples, divided by the pass\u0027s 26089 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "pgvector",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:25:06.749Z",
        "endUtc": "2026-10-06T16:25:26.776Z",
        "seconds": 20.027,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 1,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3492,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:24:48.745Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T16:23:48.745Z",
        "outsideLoadAtStart": 0.201,
        "outsideLoadRecentAtStart": 0.192,
        "outsideLoadDuring": 0.238,
        "accountingMismatch": false,
        "searches": 78204,
        "clientCpuSeconds": 41.7088,
        "clientCpuMsPerSearch": 0.5333,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:25:06.749Z to 2026-10-06T16:25:26.751Z, 80 samples, divided by the pass\u0027s 78204 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "pgvector",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:25:44.779Z",
        "endUtc": "2026-10-06T16:26:04.801Z",
        "seconds": 20.022,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 1,
            "min": 3488,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 2,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3489,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3489,
            "median": 3492,
            "max": 3497
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:25:26.777Z",
        "quietCheckLeadSeconds": 18.002,
        "outsideWindowFromUtc": "2026-10-06T16:24:26.777Z",
        "outsideLoadAtStart": 0.234,
        "outsideLoadRecentAtStart": 0.289,
        "outsideLoadDuring": 0.211,
        "accountingMismatch": false,
        "searches": 22348,
        "clientCpuSeconds": 14.9458,
        "clientCpuMsPerSearch": 0.6688,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:25:44.779Z to 2026-10-06T16:26:04.780Z, 80 samples, divided by the pass\u0027s 22348 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "weaviate",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:27:46.811Z",
        "endUtc": "2026-10-06T16:28:06.815Z",
        "seconds": 20.004,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3491,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 1,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3487,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 3,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3490,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3488,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3496
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:27:28.804Z",
        "quietCheckLeadSeconds": 18.007,
        "outsideWindowFromUtc": "2026-10-06T16:26:28.804Z",
        "outsideLoadAtStart": 0.219,
        "outsideLoadRecentAtStart": 0.184,
        "outsideLoadDuring": 0.212,
        "accountingMismatch": false,
        "searches": 3309,
        "clientCpuSeconds": 1.9696,
        "clientCpuMsPerSearch": 0.5952,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:27:46.811Z to 2026-10-06T16:28:06.812Z, 80 samples, divided by the pass\u0027s 3309 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "weaviate",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:28:24.852Z",
        "endUtc": "2026-10-06T16:28:44.864Z",
        "seconds": 20.011,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3489,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 1,
            "min": 3490,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3488,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 5,
            "min": 3406,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:28:06.816Z",
        "quietCheckLeadSeconds": 18.037,
        "outsideWindowFromUtc": "2026-10-06T16:27:06.816Z",
        "outsideLoadAtStart": 0.209,
        "outsideLoadRecentAtStart": 0.2,
        "outsideLoadDuring": 0.214,
        "accountingMismatch": false,
        "searches": 6111,
        "clientCpuSeconds": 3.8046,
        "clientCpuMsPerSearch": 0.6226,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:28:24.852Z to 2026-10-06T16:28:44.862Z, 80 samples, divided by the pass\u0027s 6111 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "clickhouse",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:31:35.610Z",
        "endUtc": "2026-10-06T16:31:55.621Z",
        "seconds": 20.011,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3463,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 1,
            "min": 3491,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3487,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3529
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:31:20.599Z",
        "quietCheckLeadSeconds": 15.011,
        "outsideWindowFromUtc": "2026-10-06T16:30:20.599Z",
        "outsideLoadAtStart": 0.231,
        "outsideLoadRecentAtStart": 0.191,
        "outsideLoadDuring": 0.213,
        "accountingMismatch": false,
        "searches": 8052,
        "clientCpuSeconds": 8.5545,
        "clientCpuMsPerSearch": 1.0624,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:31:35.610Z to 2026-10-06T16:31:55.620Z, 80 samples, divided by the pass\u0027s 8052 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "clickhouse",
        "pass": "exact",
        "startUtc": "2026-10-06T16:33:01.638Z",
        "endUtc": "2026-10-06T16:34:01.651Z",
        "seconds": 60.012,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3488,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 1,
            "min": 3476,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3486,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 3,
            "min": 3488,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 4,
            "min": 3485,
            "median": 3492,
            "max": 3507
          },
          {
            "cpu": 5,
            "min": 3400,
            "median": 3492,
            "max": 3502
          },
          {
            "cpu": 6,
            "min": 3487,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 7,
            "min": 3487,
            "median": 3492,
            "max": 3499
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:32:46.637Z",
        "quietCheckLeadSeconds": 15.001,
        "outsideWindowFromUtc": "2026-10-06T16:31:46.637Z",
        "outsideLoadAtStart": 0.244,
        "outsideLoadRecentAtStart": 0.203,
        "outsideLoadDuring": 0.244,
        "accountingMismatch": false,
        "searches": 8752,
        "clientCpuSeconds": 7.781,
        "clientCpuMsPerSearch": 0.8891,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:33:01.638Z to 2026-10-06T16:34:01.644Z, 240 samples, divided by the pass\u0027s 8752 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "clickhouse",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:34:19.658Z",
        "endUtc": "2026-10-06T16:34:39.663Z",
        "seconds": 20.005,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3482,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 1,
            "min": 3489,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 2,
            "min": 3491,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 3,
            "min": 3491,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 4,
            "min": 3487,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3489,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 6,
            "min": 3491,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3493
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:34:01.651Z",
        "quietCheckLeadSeconds": 18.007,
        "outsideWindowFromUtc": "2026-10-06T16:33:01.651Z",
        "outsideLoadAtStart": 0.244,
        "outsideLoadRecentAtStart": 0.209,
        "outsideLoadDuring": 0.244,
        "accountingMismatch": false,
        "searches": 3862,
        "clientCpuSeconds": 3.4201,
        "clientCpuMsPerSearch": 0.8856,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:34:19.658Z to 2026-10-06T16:34:39.660Z, 80 samples, divided by the pass\u0027s 3862 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "oracle",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:37:41.563Z",
        "endUtc": "2026-10-06T16:38:01.665Z",
        "seconds": 20.102,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 1196,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 1,
            "min": 1218,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 1197,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 1247,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 1342,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 1197,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 1197,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 1197,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:37:26.561Z",
        "quietCheckLeadSeconds": 15.002,
        "outsideWindowFromUtc": "2026-10-06T16:36:26.561Z",
        "outsideLoadAtStart": 0.24,
        "outsideLoadRecentAtStart": 0.248,
        "outsideLoadDuring": 0.24,
        "accountingMismatch": false,
        "searches": 56306,
        "clientCpuSeconds": 36.2184,
        "clientCpuMsPerSearch": 0.6432,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:37:41.563Z to 2026-10-06T16:38:01.647Z, 80 samples, divided by the pass\u0027s 56306 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "oracle",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:38:19.669Z",
        "endUtc": "2026-10-06T16:38:39.690Z",
        "seconds": 20.021,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 81,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3491,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 1,
            "min": 3488,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 2,
            "min": 3488,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 3,
            "min": 3490,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 6,
            "min": 3490,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 7,
            "min": 3488,
            "median": 3492,
            "max": 3502
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:38:01.666Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T16:37:01.666Z",
        "outsideLoadAtStart": 0.239,
        "outsideLoadRecentAtStart": 0.213,
        "outsideLoadDuring": 0.221,
        "accountingMismatch": false,
        "searches": 21108,
        "clientCpuSeconds": 18.7109,
        "clientCpuMsPerSearch": 0.8864,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:38:19.669Z to 2026-10-06T16:38:39.670Z, 80 samples, divided by the pass\u0027s 21108 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "oracle",
        "pass": "exact",
        "startUtc": "2026-10-06T16:38:57.692Z",
        "endUtc": "2026-10-06T16:39:57.713Z",
        "seconds": 60.022,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 1,
            "min": 3469,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3487,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 3,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 2823,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3463,
            "median": 3492,
            "max": 3502
          },
          {
            "cpu": 6,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3489,
            "median": 3492,
            "max": 3499
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:38:39.690Z",
        "quietCheckLeadSeconds": 18.002,
        "outsideWindowFromUtc": "2026-10-06T16:37:39.690Z",
        "outsideLoadAtStart": 0.23,
        "outsideLoadRecentAtStart": 0.2,
        "outsideLoadDuring": 0.209,
        "accountingMismatch": false,
        "searches": 23348,
        "clientCpuSeconds": 22.1571,
        "clientCpuMsPerSearch": 0.949,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:38:57.692Z to 2026-10-06T16:39:57.694Z, 240 samples, divided by the pass\u0027s 23348 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "sql",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:41:30.293Z",
        "endUtc": "2026-10-06T16:41:50.300Z",
        "seconds": 20.007,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3489,
            "median": 3492,
            "max": 3504
          },
          {
            "cpu": 1,
            "min": 3491,
            "median": 3492,
            "max": 3502
          },
          {
            "cpu": 2,
            "min": 3491,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 3,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3489,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 5,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3489,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:41:12.291Z",
        "quietCheckLeadSeconds": 18.003,
        "outsideWindowFromUtc": "2026-10-06T16:40:12.291Z",
        "outsideLoadAtStart": 0.219,
        "outsideLoadRecentAtStart": 0.25,
        "outsideLoadDuring": 0.204,
        "accountingMismatch": false,
        "searches": 4821,
        "clientCpuSeconds": 5.8258,
        "clientCpuMsPerSearch": 1.2084,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:41:30.293Z to 2026-10-06T16:41:50.296Z, 80 samples, divided by the pass\u0027s 4821 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "sql",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:42:08.314Z",
        "endUtc": "2026-10-06T16:42:28.327Z",
        "seconds": 20.013,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3488,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 1,
            "min": 3491,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3490,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:41:50.301Z",
        "quietCheckLeadSeconds": 18.013,
        "outsideWindowFromUtc": "2026-10-06T16:40:50.301Z",
        "outsideLoadAtStart": 0.217,
        "outsideLoadRecentAtStart": 0.214,
        "outsideLoadDuring": 0.241,
        "accountingMismatch": false,
        "searches": 12279,
        "clientCpuSeconds": 14.6373,
        "clientCpuMsPerSearch": 1.1921,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:42:08.314Z to 2026-10-06T16:42:28.325Z, 80 samples, divided by the pass\u0027s 12279 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "opensearch",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:44:48.389Z",
        "endUtc": "2026-10-06T16:45:08.398Z",
        "seconds": 20.01,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 1,
            "min": 3491,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3490,
            "median": 3492,
            "max": 3502
          },
          {
            "cpu": 5,
            "min": 3490,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:44:30.381Z",
        "quietCheckLeadSeconds": 18.008,
        "outsideWindowFromUtc": "2026-10-06T16:43:30.381Z",
        "outsideLoadAtStart": 0.208,
        "outsideLoadRecentAtStart": 0.229,
        "outsideLoadDuring": 0.2,
        "accountingMismatch": false,
        "searches": 29534,
        "clientCpuSeconds": 15.0728,
        "clientCpuMsPerSearch": 0.5104,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:44:48.389Z to 2026-10-06T16:45:08.391Z, 80 samples, divided by the pass\u0027s 29534 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "opensearch",
        "pass": "exact",
        "startUtc": "2026-10-06T16:45:26.403Z",
        "endUtc": "2026-10-06T16:46:26.423Z",
        "seconds": 60.021,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 241,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3480,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 1,
            "min": 3486,
            "median": 3492,
            "max": 3503
          },
          {
            "cpu": 2,
            "min": 3482,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 3,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3484,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3490,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3488,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:45:08.399Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T16:44:08.399Z",
        "outsideLoadAtStart": 0.201,
        "outsideLoadRecentAtStart": 0.254,
        "outsideLoadDuring": 0.209,
        "accountingMismatch": false,
        "searches": 22550,
        "clientCpuSeconds": 10.6492,
        "clientCpuMsPerSearch": 0.4722,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:45:26.403Z to 2026-10-06T16:46:26.405Z, 240 samples, divided by the pass\u0027s 22550 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "opensearch",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:46:44.426Z",
        "endUtc": "2026-10-06T16:47:04.436Z",
        "seconds": 20.01,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3447,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 1,
            "min": 3437,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3489,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3490,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3430,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3491,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 7,
            "min": 3487,
            "median": 3492,
            "max": 3494
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:46:26.424Z",
        "quietCheckLeadSeconds": 18.002,
        "outsideWindowFromUtc": "2026-10-06T16:45:26.424Z",
        "outsideLoadAtStart": 0.209,
        "outsideLoadRecentAtStart": 0.237,
        "outsideLoadDuring": 0.205,
        "accountingMismatch": false,
        "searches": 9675,
        "clientCpuSeconds": 4.499,
        "clientCpuMsPerSearch": 0.465,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:46:44.426Z to 2026-10-06T16:47:04.428Z, 80 samples, divided by the pass\u0027s 9675 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "redis",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:50:00.441Z",
        "endUtc": "2026-10-06T16:50:20.476Z",
        "seconds": 20.035,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 1,
            "min": 3491,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 2,
            "min": 3491,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 3,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 6,
            "min": 3489,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 7,
            "min": 3488,
            "median": 3492,
            "max": 3497
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:49:45.439Z",
        "quietCheckLeadSeconds": 15.002,
        "outsideWindowFromUtc": "2026-10-06T16:48:45.439Z",
        "outsideLoadAtStart": 0.196,
        "outsideLoadRecentAtStart": 0.183,
        "outsideLoadDuring": 0.197,
        "accountingMismatch": false,
        "searches": 100763,
        "clientCpuSeconds": 37.1953,
        "clientCpuMsPerSearch": 0.3691,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:50:00.441Z to 2026-10-06T16:50:20.443Z, 80 samples, divided by the pass\u0027s 100763 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "redis",
        "pass": "exact",
        "startUtc": "2026-10-06T16:51:26.498Z",
        "endUtc": "2026-10-06T16:52:26.663Z",
        "seconds": 60.165,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3489,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 1,
            "min": 3488,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 2,
            "min": 3489,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 3,
            "min": 3490,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3488,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 5,
            "min": 3490,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 6,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3488,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:51:11.494Z",
        "quietCheckLeadSeconds": 15.004,
        "outsideWindowFromUtc": "2026-10-06T16:50:11.494Z",
        "outsideLoadAtStart": 0.187,
        "outsideLoadRecentAtStart": 0.242,
        "outsideLoadDuring": 0.202,
        "accountingMismatch": false,
        "searches": 160015,
        "clientCpuSeconds": 90.6583,
        "clientCpuMsPerSearch": 0.5666,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:51:26.498Z to 2026-10-06T16:52:26.498Z, 240 samples, divided by the pass\u0027s 160015 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "redis",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:53:32.683Z",
        "endUtc": "2026-10-06T16:53:52.737Z",
        "seconds": 20.054,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3488,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 1,
            "min": 3407,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 2,
            "min": 3491,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 3,
            "min": 3489,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 6,
            "min": 3487,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 7,
            "min": 3490,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:53:17.679Z",
        "quietCheckLeadSeconds": 15.004,
        "outsideWindowFromUtc": "2026-10-06T16:52:17.679Z",
        "outsideLoadAtStart": 0.193,
        "outsideLoadRecentAtStart": 0.152,
        "outsideLoadDuring": 0.192,
        "accountingMismatch": false,
        "searches": 50378,
        "clientCpuSeconds": 28.6471,
        "clientCpuMsPerSearch": 0.5686,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:53:32.683Z to 2026-10-06T16:53:52.683Z, 80 samples, divided by the pass\u0027s 50378 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "chroma",
        "pass": "default@8",
        "startUtc": "2026-10-06T16:55:19.991Z",
        "endUtc": "2026-10-06T16:55:40.002Z",
        "seconds": 20.011,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 1,
            "min": 3490,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3489,
            "median": 3492,
            "max": 3499
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:55:01.976Z",
        "quietCheckLeadSeconds": 18.015,
        "outsideWindowFromUtc": "2026-10-06T16:54:01.976Z",
        "outsideLoadAtStart": 0.16,
        "outsideLoadRecentAtStart": 0.146,
        "outsideLoadDuring": 0.151,
        "accountingMismatch": false,
        "searches": 15548,
        "clientCpuSeconds": 7.3365,
        "clientCpuMsPerSearch": 0.4719,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:55:19.991Z to 2026-10-06T16:55:39.998Z, 80 samples, divided by the pass\u0027s 15548 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "chroma",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:55:58.007Z",
        "endUtc": "2026-10-06T16:56:18.015Z",
        "seconds": 20.008,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3489,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 1,
            "min": 3489,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3492,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:55:40.003Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T16:54:40.003Z",
        "outsideLoadAtStart": 0.152,
        "outsideLoadRecentAtStart": 0.124,
        "outsideLoadDuring": 0.152,
        "accountingMismatch": false,
        "searches": 9377,
        "clientCpuSeconds": 4.4669,
        "clientCpuMsPerSearch": 0.4764,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:55:58.007Z to 2026-10-06T16:56:18.007Z, 80 samples, divided by the pass\u0027s 9377 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "typesense",
        "pass": "exact",
        "startUtc": "2026-10-06T16:58:20.012Z",
        "endUtc": "2026-10-06T16:59:20.021Z",
        "seconds": 60.009,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 1,
            "min": 3488,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 2,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3489,
            "median": 3492,
            "max": 3502
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3434,
            "median": 3492,
            "max": 3503
          },
          {
            "cpu": 6,
            "min": 3491,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 7,
            "min": 3475,
            "median": 3492,
            "max": 3500
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:58:02.004Z",
        "quietCheckLeadSeconds": 18.008,
        "outsideWindowFromUtc": "2026-10-06T16:57:02.004Z",
        "outsideLoadAtStart": 0.139,
        "outsideLoadRecentAtStart": 0.052,
        "outsideLoadDuring": 0.213,
        "accountingMismatch": false,
        "searches": 10446,
        "clientCpuSeconds": 5.6438,
        "clientCpuMsPerSearch": 0.5403,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:58:20.012Z to 2026-10-06T16:59:20.013Z, 240 samples, divided by the pass\u0027s 10446 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "typesense",
        "pass": "default@1",
        "startUtc": "2026-10-06T16:59:38.026Z",
        "endUtc": "2026-10-06T16:59:58.032Z",
        "seconds": 20.006,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3488,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 1,
            "min": 3489,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 2,
            "min": 3490,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3489,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3489,
            "median": 3492,
            "max": 3501
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:59:20.022Z",
        "quietCheckLeadSeconds": 18.004,
        "outsideWindowFromUtc": "2026-10-06T16:58:20.022Z",
        "outsideLoadAtStart": 0.213,
        "outsideLoadRecentAtStart": 0.212,
        "outsideLoadDuring": 0.202,
        "accountingMismatch": false,
        "searches": 4716,
        "clientCpuSeconds": 2.456,
        "clientCpuMsPerSearch": 0.5208,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T16:59:38.026Z to 2026-10-06T16:59:58.028Z, 80 samples, divided by the pass\u0027s 4716 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "typesense",
        "pass": "default@8",
        "startUtc": "2026-10-06T17:00:16.050Z",
        "endUtc": "2026-10-06T17:00:36.060Z",
        "seconds": 20.011,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 1,
            "min": 3489,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3490,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 5,
            "min": 3490,
            "median": 3492,
            "max": 3505
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T16:59:58.033Z",
        "quietCheckLeadSeconds": 18.017,
        "outsideWindowFromUtc": "2026-10-06T16:58:58.033Z",
        "outsideLoadAtStart": 0.205,
        "outsideLoadRecentAtStart": 0.188,
        "outsideLoadDuring": 0.049,
        "accountingMismatch": false,
        "searches": 11527,
        "clientCpuSeconds": 6.8218,
        "clientCpuMsPerSearch": 0.5918,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T17:00:16.050Z to 2026-10-06T17:00:36.057Z, 80 samples, divided by the pass\u0027s 11527 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "mongodb",
        "pass": "default@1",
        "startUtc": "2026-10-06T17:02:47.896Z",
        "endUtc": "2026-10-06T17:03:07.912Z",
        "seconds": 20.015,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 81,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3491,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 1,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3488,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3490,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3492,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 6,
            "min": 3491,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3498
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T17:02:29.893Z",
        "quietCheckLeadSeconds": 18.003,
        "outsideWindowFromUtc": "2026-10-06T17:01:29.893Z",
        "outsideLoadAtStart": 0.266,
        "outsideLoadRecentAtStart": 0.291,
        "outsideLoadDuring": 0.35,
        "accountingMismatch": false,
        "searches": 15409,
        "clientCpuSeconds": 4.6703,
        "clientCpuMsPerSearch": 0.3031,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T17:02:47.896Z to 2026-10-06T17:03:07.897Z, 80 samples, divided by the pass\u0027s 15409 completed searches",
        "waitedSeconds": 1,
        "busyBox": true,
        "busyReason": "processes outside the benchmark used 0.35 CPUs on average during the pass, above the limit of 0.3"
      },
      {
        "target": "mongodb",
        "pass": "default@8",
        "startUtc": "2026-10-06T17:03:36.920Z",
        "endUtc": "2026-10-06T17:03:56.936Z",
        "seconds": 20.016,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 1,
            "min": 3491,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 2,
            "min": 3490,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3488,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3503
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3495
          },
          {
            "cpu": 6,
            "min": 3490,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3486,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T17:03:18.914Z",
        "quietCheckLeadSeconds": 18.005,
        "outsideWindowFromUtc": "2026-10-06T17:02:18.914Z",
        "outsideLoadAtStart": 0.299,
        "outsideLoadRecentAtStart": 0.215,
        "outsideLoadDuring": 0.251,
        "accountingMismatch": false,
        "searches": 47043,
        "clientCpuSeconds": 13.6976,
        "clientCpuMsPerSearch": 0.2912,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T17:03:36.920Z to 2026-10-06T17:03:56.922Z, 79 samples, divided by the pass\u0027s 47043 completed searches",
        "waitedSeconds": 11,
        "busyBox": false
      },
      {
        "target": "mongodb",
        "pass": "exact",
        "startUtc": "2026-10-06T17:04:14.939Z",
        "endUtc": "2026-10-06T17:05:14.981Z",
        "seconds": 60.042,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 240,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 1,
            "min": 3365,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 2,
            "min": 3441,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 3,
            "min": 3490,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3501
          },
          {
            "cpu": 6,
            "min": 3390,
            "median": 3492,
            "max": 3500
          },
          {
            "cpu": 7,
            "min": 3489,
            "median": 3492,
            "max": 3499
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T17:03:56.936Z",
        "quietCheckLeadSeconds": 18.003,
        "outsideWindowFromUtc": "2026-10-06T17:02:56.936Z",
        "outsideLoadAtStart": 0.244,
        "outsideLoadRecentAtStart": 0.242,
        "outsideLoadDuring": 0.301,
        "accountingMismatch": false,
        "searches": 47194,
        "clientCpuSeconds": 14.2413,
        "clientCpuMsPerSearch": 0.3018,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T17:04:14.939Z to 2026-10-06T17:05:14.940Z, 240 samples, divided by the pass\u0027s 47194 completed searches",
        "waitedSeconds": 0,
        "busyBox": true,
        "busyReason": "processes outside the benchmark used 0.30 CPUs on average during the pass, above the limit of 0.3"
      },
      {
        "target": "milvus",
        "pass": "default@1",
        "startUtc": "2026-10-06T17:10:03.627Z",
        "endUtc": "2026-10-06T17:10:23.633Z",
        "seconds": 20.007,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3491,
            "median": 3492,
            "max": 3494
          },
          {
            "cpu": 1,
            "min": 3490,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 2,
            "min": 3491,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3497
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3493
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3493
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T17:09:45.625Z",
        "quietCheckLeadSeconds": 18.002,
        "outsideWindowFromUtc": "2026-10-06T17:08:45.625Z",
        "outsideLoadAtStart": 0.219,
        "outsideLoadRecentAtStart": 0.222,
        "outsideLoadDuring": 0.197,
        "accountingMismatch": false,
        "searches": 8145,
        "clientCpuSeconds": 4.632,
        "clientCpuMsPerSearch": 0.5687,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T17:10:03.627Z to 2026-10-06T17:10:23.627Z, 80 samples, divided by the pass\u0027s 8145 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      },
      {
        "target": "milvus",
        "pass": "default@8",
        "startUtc": "2026-10-06T17:10:41.641Z",
        "endUtc": "2026-10-06T17:11:01.649Z",
        "seconds": 20.008,
        "endedBy": "result line",
        "governor": "performance",
        "frequencySamples": 80,
        "nearestSample": false,
        "engineMhzMedian": 3492,
        "clientMhzMedian": 3492,
        "cpuMhz": [
          {
            "cpu": 0,
            "min": 3490,
            "median": 3492,
            "max": 3496
          },
          {
            "cpu": 1,
            "min": 3492,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 2,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 3,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 4,
            "min": 3491,
            "median": 3492,
            "max": 3498
          },
          {
            "cpu": 5,
            "min": 3491,
            "median": 3492,
            "max": 3516
          },
          {
            "cpu": 6,
            "min": 3492,
            "median": 3492,
            "max": 3492
          },
          {
            "cpu": 7,
            "min": 3492,
            "median": 3492,
            "max": 3492
          }
        ],
        "quietCheckBefore": "warm-up",
        "quietCheckUtc": "2026-10-06T17:10:23.634Z",
        "quietCheckLeadSeconds": 18.007,
        "outsideWindowFromUtc": "2026-10-06T17:09:23.634Z",
        "outsideLoadAtStart": 0.214,
        "outsideLoadRecentAtStart": 0.17,
        "outsideLoadDuring": 0.249,
        "accountingMismatch": false,
        "searches": 25258,
        "clientCpuSeconds": 15.8831,
        "clientCpuMsPerSearch": 0.6288,
        "clientCpuBasis": "user \u002B system CPU time of this process\u0027s threads (not its finished children; includes the 4 Hz sampler thread and the kernel network time spent on this process\u0027s behalf) from 2026-10-06T17:10:41.641Z to 2026-10-06T17:11:01.643Z, 80 samples, divided by the pass\u0027s 25258 completed searches",
        "waitedSeconds": 0,
        "busyBox": false
      }
    ],
    "cpuIdle": {
      "driver": "intel_idle",
      "governor": "menu",
      "intelIdleMaxCstate": "9",
      "states": [
        {
          "index": 0,
          "name": "POLL",
          "latencyUs": 0,
          "residencyUs": 0,
          "disabledOn": "",
          "defaultStatus": "enabled"
        },
        {
          "index": 1,
          "name": "C1",
          "latencyUs": 2,
          "residencyUs": 2,
          "disabledOn": "",
          "defaultStatus": "enabled"
        },
        {
          "index": 2,
          "name": "C1E",
          "latencyUs": 10,
          "residencyUs": 20,
          "disabledOn": "",
          "defaultStatus": "enabled"
        },
        {
          "index": 3,
          "name": "C3",
          "latencyUs": 33,
          "residencyUs": 100,
          "disabledOn": "0-7",
          "defaultStatus": "disabled"
        },
        {
          "index": 4,
          "name": "C6",
          "latencyUs": 133,
          "residencyUs": 400,
          "disabledOn": "",
          "defaultStatus": "enabled"
        }
      ]
    },
    "cpuIdleAtEnd": {
      "driver": "intel_idle",
      "governor": "menu",
      "intelIdleMaxCstate": "9",
      "states": [
        {
          "index": 0,
          "name": "POLL",
          "latencyUs": 0,
          "residencyUs": 0,
          "disabledOn": "",
          "defaultStatus": "enabled"
        },
        {
          "index": 1,
          "name": "C1",
          "latencyUs": 2,
          "residencyUs": 2,
          "disabledOn": "",
          "defaultStatus": "enabled"
        },
        {
          "index": 2,
          "name": "C1E",
          "latencyUs": 10,
          "residencyUs": 20,
          "disabledOn": "",
          "defaultStatus": "enabled"
        },
        {
          "index": 3,
          "name": "C3",
          "latencyUs": 33,
          "residencyUs": 100,
          "disabledOn": "0-7",
          "defaultStatus": "disabled"
        },
        {
          "index": 4,
          "name": "C6",
          "latencyUs": 133,
          "residencyUs": 400,
          "disabledOn": "",
          "defaultStatus": "enabled"
        }
      ]
    },
    "connections": [
      {
        "target": "sqlitevec",
        "hosting": "embedded",
        "route": "in this process (embedded), no network",
        "endpoints": [],
        "observed": [
          {
            "address": "localhost",
            "port": 1433,
            "sockets": 2,
            "kind": "not this target"
          }
        ],
        "observedUtc": "2026-10-06T15:52:43.630Z"
      },
      {
        "target": "mariadb",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvbbench-mariadb",
            "address": "container-ip",
            "port": 3306,
            "containerId": "1ff5bbe73276",
            "network": "gvbbench-mariadb_default",
            "notUsed": "localhost:13306"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 3306,
            "sockets": 8,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T15:56:55.560Z"
      },
      {
        "target": "duckdb",
        "hosting": "embedded",
        "route": "in this process (embedded), no network",
        "endpoints": [],
        "observed": [],
        "observedUtc": "2026-10-06T16:01:01.822Z",
        "problem": "no open TCP connection was found"
      },
      {
        "target": "qdrant",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-qdrant",
            "address": "container-ip",
            "port": 6334,
            "containerId": "e08d8b70ccba",
            "network": "gvb-qdrant_default",
            "notUsed": "localhost:16334"
          },
          {
            "what": "container gvb-qdrant",
            "address": "container-ip",
            "port": 6333,
            "containerId": "e08d8b70ccba",
            "network": "gvb-qdrant_default",
            "notUsed": "localhost:16333"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 6334,
            "sockets": 1,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:03:25.766Z"
      },
      {
        "target": "elasticsearch",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-elasticsearch",
            "address": "container-ip",
            "port": 9200,
            "containerId": "2fe4965f42f7",
            "network": "engines_default",
            "notUsed": "localhost:9200"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 9200,
            "sockets": 1,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:08:04.903Z"
      },
      {
        "target": "vespa",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-vespa",
            "address": "container-ip",
            "port": 8080,
            "containerId": "862c328d26b8",
            "network": "engines_default",
            "notUsed": "localhost:8090"
          },
          {
            "what": "container gvb-vespa",
            "address": "container-ip",
            "port": 19071,
            "containerId": "862c328d26b8",
            "network": "engines_default",
            "notUsed": "localhost:19071"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 8080,
            "sockets": 1,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:13:08.909Z"
      },
      {
        "target": "qdrant-hnsw",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-qdrant",
            "address": "container-ip",
            "port": 6334,
            "containerId": "fa2ccc8859f4",
            "network": "gvb-qdrant_default",
            "notUsed": "localhost:16334"
          },
          {
            "what": "container gvb-qdrant",
            "address": "container-ip",
            "port": 6333,
            "containerId": "fa2ccc8859f4",
            "network": "gvb-qdrant_default",
            "notUsed": "localhost:16333"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 6334,
            "sockets": 1,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:17:31.069Z"
      },
      {
        "target": "sql-diskann",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-mssql",
            "address": "container-ip",
            "port": 1433,
            "containerId": "8b2960fc1293",
            "network": "gvb-mssql_default",
            "notUsed": "localhost:14330"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 1433,
            "sockets": 9,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:21:51.254Z"
      },
      {
        "target": "pgvector",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-pgvector",
            "address": "container-ip",
            "port": 5432,
            "containerId": "a107b4cc5b1b",
            "network": "gvb-pgvector_default",
            "notUsed": "localhost:5432"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 5432,
            "sockets": 8,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:26:04.801Z"
      },
      {
        "target": "weaviate",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-weaviate",
            "address": "container-ip",
            "port": 8080,
            "containerId": "f60c5682a178",
            "network": "gvb-weaviate_default",
            "notUsed": "localhost:8085"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 8080,
            "sockets": 8,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:28:44.864Z"
      },
      {
        "target": "clickhouse",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-clickhouse",
            "address": "container-ip",
            "port": 8123,
            "containerId": "1b38cf38a4a4",
            "network": "gvb-clickhouse_default",
            "notUsed": "localhost:8123"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 8123,
            "sockets": 1,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:34:39.663Z"
      },
      {
        "target": "oracle",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-oracle",
            "address": "container-ip",
            "port": 1521,
            "containerId": "7900917130ab",
            "network": "gvb-oracle_default",
            "notUsed": "localhost:1521"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 1521,
            "sockets": 14,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:39:57.714Z"
      },
      {
        "target": "sql",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-mssql",
            "address": "container-ip",
            "port": 1433,
            "containerId": "703c36ce4875",
            "network": "gvb-mssql_default",
            "notUsed": "localhost:14330"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 1433,
            "sockets": 9,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:42:28.327Z"
      },
      {
        "target": "opensearch",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-opensearch",
            "address": "container-ip",
            "port": 9200,
            "containerId": "aa945bc1d1e9",
            "network": "engines_default",
            "notUsed": "localhost:9201"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 9200,
            "sockets": 1,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:47:04.436Z"
      },
      {
        "target": "redis",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-redis",
            "address": "container-ip",
            "port": 6379,
            "containerId": "78e85925c519",
            "network": "gvb-redis_default",
            "notUsed": "localhost:6379"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 6379,
            "sockets": 1,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:53:52.737Z"
      },
      {
        "target": "chroma",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-chroma",
            "address": "container-ip",
            "port": 8000,
            "containerId": "dd4da1e51118",
            "network": "gvb-chroma_default",
            "notUsed": "localhost:8000"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 8000,
            "sockets": 8,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T16:56:18.015Z"
      },
      {
        "target": "typesense",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-typesense",
            "address": "container-ip",
            "port": 8108,
            "containerId": "8ca3b00f385a",
            "network": "engines_default",
            "notUsed": "localhost:8108"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 8108,
            "sockets": 8,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T17:00:36.060Z"
      },
      {
        "target": "mongodb",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-mongodb",
            "address": "container-ip",
            "port": 27017,
            "containerId": "30db26986ada",
            "network": "gvb-mongodb_default",
            "notUsed": "localhost:27017"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 27017,
            "sockets": 10,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T17:05:14.981Z"
      },
      {
        "target": "milvus",
        "hosting": "compose",
        "route": "container address on its Docker network, not the published port (no docker-proxy)",
        "endpoints": [
          {
            "what": "container gvb-milvus",
            "address": "container-ip",
            "port": 19530,
            "containerId": "6a4ba4e1d723",
            "network": "gvb-milvus_default",
            "notUsed": "localhost:19530"
          },
          {
            "what": "container gvb-milvus",
            "address": "container-ip",
            "port": 9091,
            "containerId": "6a4ba4e1d723",
            "network": "gvb-milvus_default",
            "notUsed": "localhost:9091"
          }
        ],
        "observed": [
          {
            "address": "container-ip",
            "port": 19530,
            "sockets": 8,
            "kind": "this target"
          }
        ],
        "observedUtc": "2026-10-06T17:11:01.649Z"
      }
    ],
    "cpuAccounting": "busy = user \u002B nice \u002B system \u002B irq \u002B softirq \u002B steal (guest time is already in user); CONFIG_IRQ_TIME_ACCOUNTING is not set, so task and cgroup run time include the interrupt and softirq time that hit them and it is added; CONFIG_PARAVIRT_TIME_ACCOUNTING is not set, so steal is added (/boot/config-6.8.0-142-generic)",
    "loadAverageEnd": "4.12 2.65 2.76",
    "stateFile": "~/gvb-work/bench-machine-state.json",
    "restored": [
      "client affinity back to 0-7",
      "governor back to schedutil on CPUs 0-7",
      "CPU clock back: turbo on (/sys/devices/system/cpu/intel_pstate/no_turbo 0); every CPU\u0027s ceiling and floor read as before the run; uncore limit back to min 1200 MHz, max 3000 MHz (MSR 0x620 = 0xc1e)"
    ],
    "restoreProblems": [],
    "flags": [
      "sqlitevec is embedded: it ran inside the client process on the client CPUs 0-1,4-5, sharing them with the client.",
      "duckdb is embedded: it ran inside the client process on the client CPUs 0-1,4-5, sharing them with the client.",
      "WARNING: busy box during mongodb default@1: processes outside the benchmark used 0.35 CPUs on average during the pass, above the limit of 0.3.",
      "WARNING: busy box during mongodb exact: processes outside the benchmark used 0.30 CPUs on average during the pass, above the limit of 0.3.",
      "WARNING: clock off its pinned value during oracle default@8: the engine CPUs averaged 3168 MHz, -9.5% against the pinned 3500 MHz; the client CPUs averaged 3214 MHz, -8.2% against the pinned 3500 MHz (limit 1%); this pass is not comparable with passes at the pinned clock."
    ]
  }
}