adriangbot commented on PR #24086: URL: https://github.com/apache/datafusion/pull/24086#issuecomment-5852584767
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/24086#issuecomment-5852458219) **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) Comparing claude/datafusion-rowgroup-buffering-ec9abc (4a9f276c58c8b1d24b5bb64a182a5658ebb1bdfb) to e8e41ae (merge-base) [diff](https://github.com/apache/datafusion/compare/e8e41ae958a5e6856ab9b0e7d8fb5deece0a9e91..4a9f276c58c8b1d24b5bb64a182a5658ebb1bdfb) <details><summary>Run configuration</summary> ```yaml run benchmark tpch env: DATAFUSION_EXECUTION_PARQUET_PUSHDOWN_FILTERS: "true" DATAFUSION_RUNTIME_MEMORY_LIMIT: "1G" SIMULATE_LATENCY: "true" changed: env: DATAFUSION_EXECUTION_PARQUET_READ_AHEAD_BYTES: "104857600" ``` </details> <details><summary>CPU Details (lscpu)</summary> ``` Architecture: aarch64 CPU op-mode(s): 64-bit Byte Order: Little Endian CPU(s): 16 On-line CPU(s) list: 0-15 Vendor ID: ARM Model name: Neoverse-V2 Model: 1 Thread(s) per core: 1 Core(s) per cluster: 16 Socket(s): - Cluster(s): 1 Stepping: r0p1 BogoMIPS: 2000.00 Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti L1d cache: 1 MiB (16 instances) L1i cache: 1 MiB (16 instances) L2 cache: 32 MiB (16 instances) L3 cache: 80 MiB (1 instance) NUMA node(s): 1 NUMA node0 CPU(s): 0-15 Vulnerability Gather data sampling: Not affected Vulnerability Indirect target selection: Not affected Vulnerability Itlb multihit: Not affected Vulnerability L1tf: Not affected Vulnerability Mds: Not affected Vulnerability Meltdown: Not affected Vulnerability Mmio stale data: Not affected Vulnerability Reg file data sampling: Not affected Vulnerability Retbleed: Not affected Vulnerability Spec rstack overflow: Not affected Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl Vulnerability Spectre v1: Mitigation; __user pointer sanitization Vulnerability Spectre v2: Mitigation; CSV2, BHB Vulnerability Srbds: Not affected Vulnerability Tsa: Not affected Vulnerability Tsx async abort: Not affected Vulnerability Vmscape: Not affected ``` </details> <details><summary>Details</summary> <p> ``` Comparing HEAD and claude_datafusion-rowgroup-buffering-ec9abc -------------------- Benchmark tpch_sf1.json -------------------- βββββββββββββ³βββββββββββββ³ββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ β Query β HEAD β claude_datafusion-rowgroup-buffering-ec9abc β Change β β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β QQuery 1 β 887.59 ms β 215.47 ms β +4.12x faster β β QQuery 2 β 1882.57 ms β 726.73 ms β +2.59x faster β β QQuery 3 β 2050.02 ms β 514.26 ms β +3.99x faster β β QQuery 4 β 1291.41 ms β 405.90 ms β +3.18x faster β β QQuery 5 β 1766.54 ms β 522.07 ms β +3.38x faster β β QQuery 6 β 1821.60 ms β 205.43 ms β +8.87x faster β β QQuery 7 β 2700.55 ms β 608.13 ms β +4.44x faster β β QQuery 8 β 2859.58 ms β 726.92 ms β +3.93x faster β β QQuery 9 β 2478.88 ms β 855.36 ms β +2.90x faster β β QQuery 10 β 2060.22 ms β 529.39 ms β +3.89x faster β β QQuery 11 β 1321.85 ms β 622.86 ms β +2.12x faster β β QQuery 12 β 2107.50 ms β 416.37 ms β +5.06x faster β β QQuery 13 β 746.97 ms β 326.69 ms β +2.29x faster β β QQuery 14 β 1156.02 ms β 241.94 ms β +4.78x faster β β QQuery 15 β 2399.55 ms β 450.05 ms β +5.33x faster β β QQuery 16 β 648.46 ms β 263.60 ms β +2.46x faster β β QQuery 17 β 2220.75 ms β 511.71 ms β +4.34x faster β β QQuery 18 β 2187.58 ms β 821.92 ms β +2.66x faster β β QQuery 19 β 2392.12 ms β 283.83 ms β +8.43x faster β β QQuery 20 β 2068.55 ms β 445.24 ms β +4.65x faster β β QQuery 21 β 4469.31 ms β 864.37 ms β +5.17x faster β β QQuery 22 β 762.89 ms β 499.58 ms β +1.53x faster β βββββββββββββ΄βββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ β Benchmark Summary β β β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β Total Time (HEAD) β 42280.50ms β β Total Time (claude_datafusion-rowgroup-buffering-ec9abc) β 11057.81ms β β Average Time (HEAD) β 1921.84ms β β Average Time (claude_datafusion-rowgroup-buffering-ec9abc) β 502.63ms β β Queries Faster β 22 β β Queries Slower β 0 β β Queries with No Change β 0 β β Queries with Failure β 0 β ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ Distribution per query (min / mean Β±stddev / max): Comparing HEAD and claude_datafusion-rowgroup-buffering-ec9abc -------------------- Benchmark tpch_sf1.json -------------------- βββββββββββββ³βββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ β Query β HEAD β claude_datafusion-rowgroup-buffering-ec9abc β Change β β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β QQuery 1 β 887.59 / 1060.44 Β±118.95 / 1191.63 ms β 215.47 / 232.24 Β±10.71 / 243.28 ms β +4.57x faster β β QQuery 2 β 1882.57 / 1950.80 Β±72.43 / 2064.43 ms β 726.73 / 848.15 Β±78.71 / 959.48 ms β +2.30x faster β β QQuery 3 β 2050.02 / 2306.95 Β±164.27 / 2513.76 ms β 514.26 / 564.99 Β±36.40 / 615.87 ms β +4.08x faster β β QQuery 4 β 1291.41 / 1454.06 Β±122.58 / 1626.52 ms β 405.90 / 412.65 Β±8.47 / 429.12 ms β +3.52x faster β β QQuery 5 β 1766.54 / 1882.70 Β±85.96 / 2032.56 ms β 522.07 / 570.21 Β±33.47 / 611.44 ms β +3.30x faster β β QQuery 6 β 1821.60 / 1972.05 Β±110.59 / 2153.31 ms β 205.43 / 218.35 Β±10.19 / 227.74 ms β +9.03x faster β β QQuery 7 β 2700.55 / 2870.11 Β±116.14 / 3063.92 ms β 608.13 / 699.30 Β±70.24 / 801.79 ms β +4.10x faster β β QQuery 8 β 2859.58 / 3122.91 Β±163.33 / 3297.12 ms β 726.92 / 781.32 Β±67.99 / 902.42 ms β +4.00x faster β β QQuery 9 β 2478.88 / 2624.50 Β±128.08 / 2862.09 ms β 855.36 / 876.73 Β±23.64 / 909.05 ms β +2.99x faster β β QQuery 10 β 2060.22 / 2167.32 Β±125.04 / 2412.15 ms β 529.39 / 559.72 Β±26.45 / 596.98 ms β +3.87x faster β β QQuery 11 β 1321.85 / 1427.97 Β±94.58 / 1554.46 ms β 622.86 / 696.44 Β±43.35 / 756.79 ms β +2.05x faster β β QQuery 12 β 2107.50 / 2221.29 Β±99.80 / 2381.68 ms β 416.37 / 424.74 Β±9.23 / 442.02 ms β +5.23x faster β β QQuery 13 β 746.97 / 798.68 Β±43.41 / 877.38 ms β 326.69 / 371.94 Β±32.72 / 408.70 ms β +2.15x faster β β QQuery 14 β 1156.02 / 1227.91 Β±41.87 / 1274.82 ms β 241.94 / 334.44 Β±58.73 / 419.82 ms β +3.67x faster β β QQuery 15 β 2399.55 / 2522.95 Β±76.58 / 2607.35 ms β 450.05 / 549.98 Β±81.83 / 623.53 ms β +4.59x faster β β QQuery 16 β 648.46 / 769.11 Β±70.79 / 870.47 ms β 263.60 / 328.85 Β±60.60 / 399.88 ms β +2.34x faster β β QQuery 17 β 2220.75 / 2342.17 Β±108.10 / 2492.67 ms β 511.71 / 567.61 Β±38.88 / 633.14 ms β +4.13x faster β β QQuery 18 β 2187.58 / 2306.43 Β±85.98 / 2402.84 ms β 821.92 / 828.33 Β±4.11 / 833.78 ms β +2.78x faster β β QQuery 19 β 2392.12 / 2653.26 Β±162.13 / 2837.26 ms β 283.83 / 345.32 Β±40.94 / 412.55 ms β +7.68x faster β β QQuery 20 β 2068.55 / 2263.11 Β±137.36 / 2471.28 ms β 445.24 / 521.39 Β±56.65 / 616.16 ms β +4.34x faster β β QQuery 21 β 4469.31 / 4756.34 Β±169.16 / 4988.58 ms β 864.37 / 882.34 Β±11.33 / 894.66 ms β +5.39x faster β β QQuery 22 β 762.89 / 790.97 Β±17.49 / 812.32 ms β 499.58 / 501.25 Β±1.69 / 503.63 ms β +1.58x faster β βββββββββββββ΄βββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ β Benchmark Summary β β β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β Total Time (HEAD) β 45492.04ms β β Total Time (claude_datafusion-rowgroup-buffering-ec9abc) β 12116.29ms β β Average Time (HEAD) β 2067.82ms β β Average Time (claude_datafusion-rowgroup-buffering-ec9abc) β 550.74ms β β Queries Faster β 22 β β Queries Slower β 0 β β Queries with No Change β 0 β β Queries with Failure β 0 β ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ ``` </p> </details> <details><summary>Memory Pool Peaks</summary> Peak `MemoryPool` reservation per query β what DataFusion's accounting believes it reserved. Recorded only for benchmarks that write a results JSON (the `dfbench` suites, and the suites `bench.sh` runs through the Criterion SQL harness since apache/datafusion#25644), and only under a memory limit: `DATAFUSION_RUNTIME_MEMORY_LIMIT`, or one the suite sets itself. Base: `e8e41ae (merge-base)` | Changed: `claude/datafusion-rowgroup-buffering-ec9abc` **`tpch` β `tpch_sf1`** | Query | Base | Changed | Change | | --- | --- | --- | --- | | Query 1 | 30.0 MiB | 48.6 MiB | +61.8% | | Query 2 | 14.8 MiB | 14.8 MiB | +0.0% | | Query 3 | 11.5 MiB | 57.9 MiB | +404.6% | | Query 4 | 23.0 MiB | 30.3 MiB | +31.6% | | Query 5 | 32.7 MiB | 64.2 MiB | +96.4% | | Query 6 | 832 B | 44.3 MiB | +5585866.2% | | Query 7 | 99.4 MiB | 113.5 MiB | +14.2% | | Query 8 | 31.6 MiB | 102.9 MiB | +225.4% | | Query 9 | 153.9 MiB | 209.6 MiB | +36.2% | | Query 10 | 52.8 MiB | 77.3 MiB | +46.3% | | Query 11 | 30.8 MiB | 41.1 MiB | +33.7% | | Query 12 | 20.0 MiB | 38.7 MiB | +93.7% | | Query 13 | 52.6 MiB | 52.3 MiB | -0.5% | | Query 14 | 9.6 MiB | 65.8 MiB | +586.3% | | Query 15 | 12.5 MiB | 52.3 MiB | +319.3% | | Query 16 | 45.1 MiB | 34.3 MiB | -24.0% | | Query 17 | 6.7 MiB | 58.7 MiB | +775.1% | | Query 18 | 85.2 MiB | 84.1 MiB | -1.3% | | Query 19 | 187.1 KiB | 64.9 MiB | +35395.4% | | Query 20 | 10.9 MiB | 50.8 MiB | +368.2% | | Query 21 | 42.4 MiB | 78.9 MiB | +85.9% | | Query 22 | 25.3 MiB | 25.3 MiB | +0.0% | **Pool accounting vs. process RSS** Max pool peak is the largest reservation any single query in the run reached; peak RSS covers the whole invocation, including data loading and allocator retention, and the two high-water marks need not coincide in time. The gap is therefore an upper bound on what the pool did not account for, not a measurement of it. | Benchmark | Side | Max pool peak | Peak RSS | Gap | RSS / pool | | --- | --- | --- | --- | --- | --- | | tpch | base (`e8e41ae (merge-base)`) | 153.9 MiB | 641.1 MiB | 487.3 MiB | 4.2Γ | | tpch | changed (`claude/datafusion-rowgroup-buffering-ec9abc`) | 209.6 MiB | 697.3 MiB | 487.7 MiB | 3.3Γ | </details> <details><summary>Resource Usage</summary> **tpch β base (merge-base)** | Metric | Value | |--------|-------| | Wall time | 230.0s | | Peak memory | 641.1 MiB | | Avg memory | 398.5 MiB | | CPU user | 34.9s | | CPU sys | 5.1s | | Peak spill | 0 B | **tpch β branch** | Metric | Value | |--------|-------| | Wall time | 65.0s | | Peak memory | 697.3 MiB | | Avg memory | 415.1 MiB | | CPU user | 37.6s | | CPU sys | 3.4s | | Peak spill | 0 B | </details> --- [File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
