adriangbot commented on PR #23565: URL: https://github.com/apache/datafusion/pull/23565#issuecomment-5249076127
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/23565#issuecomment-5249003089) **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) Comparing spill-dedup-view-arrays (20bc3e1a140f8519ce80af05a8d446de9a505d8a) to a9b61ab (merge-base) [diff](https://github.com/apache/datafusion/compare/a9b61abf04b59ff08e7254b598b07f4fa058c5d1..20bc3e1a140f8519ce80af05a8d446de9a505d8a) <details><summary>Run configuration</summary> ```yaml run benchmark sort_tpch env: DATAFUSION_EXECUTION_TARGET_PARTITIONS: "4" DATAFUSION_RUNTIME_MEMORY_LIMIT: "512M" ``` </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 spill-dedup-view-arrays -------------------- Benchmark sort_tpch1.json -------------------- βββββββββ³βββββββββββββ³ββββββββββββββββββββββββββ³βββββββββββββββ β Query β HEAD β spill-dedup-view-arrays β Change β β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β Q1 β 150.39 ms β 151.06 ms β no change β β Q2 β 135.83 ms β 136.40 ms β no change β β Q3 β 602.63 ms β 699.14 ms β 1.16x slower β β Q4 β 249.09 ms β 252.54 ms β no change β β Q5 β 264.93 ms β 267.72 ms β no change β β Q6 β 290.46 ms β 292.75 ms β no change β β Q7 β 725.17 ms β 734.68 ms β no change β β Q8 β 503.43 ms β 528.52 ms β no change β β Q9 β 540.82 ms β 565.44 ms β no change β β Q10 β 1227.66 ms β 1263.74 ms β no change β β Q11 β 425.69 ms β 454.23 ms β 1.07x slower β βββββββββ΄βββββββββββββ΄ββββββββββββββββββββββββββ΄βββββββββββββββ ββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ β Benchmark Summary β β β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββ© β Total Time (HEAD) β 5116.11ms β β Total Time (spill-dedup-view-arrays) β 5346.22ms β β Average Time (HEAD) β 465.10ms β β Average Time (spill-dedup-view-arrays) β 486.02ms β β Queries Faster β 0 β β Queries Slower β 2 β β Queries with No Change β 9 β β Queries with Failure β 0 β ββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ Distribution per query (min / mean Β±stddev / max): Comparing HEAD and spill-dedup-view-arrays -------------------- Benchmark sort_tpch1.json -------------------- βββββββββ³βββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββββββββββββββββββββββββββ³βββββββββββββββ β Query β HEAD β spill-dedup-view-arrays β Change β β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β Q1 β 150.39 / 151.27 Β±1.12 / 153.35 ms β 151.06 / 151.79 Β±0.92 / 153.55 ms β no change β β Q2 β 135.83 / 138.62 Β±2.36 / 142.10 ms β 136.40 / 139.36 Β±3.03 / 143.18 ms β no change β β Q3 β 602.63 / 613.91 Β±6.65 / 622.24 ms β 699.14 / 706.24 Β±7.39 / 720.05 ms β 1.15x slower β β Q4 β 249.09 / 256.85 Β±5.68 / 263.50 ms β 252.54 / 257.31 Β±4.92 / 266.73 ms β no change β β Q5 β 264.93 / 268.12 Β±2.14 / 270.99 ms β 267.72 / 271.01 Β±1.86 / 273.09 ms β no change β β Q6 β 290.46 / 295.84 Β±3.90 / 301.12 ms β 292.75 / 296.51 Β±2.50 / 300.19 ms β no change β β Q7 β 725.17 / 735.03 Β±5.19 / 738.83 ms β 734.68 / 742.74 Β±5.68 / 750.06 ms β no change β β Q8 β 503.43 / 506.24 Β±2.00 / 508.36 ms β 528.52 / 532.73 Β±3.08 / 537.99 ms β 1.05x slower β β Q9 β 540.82 / 549.17 Β±7.11 / 559.09 ms β 565.44 / 572.55 Β±5.80 / 579.23 ms β no change β β Q10 β 1227.66 / 1237.65 Β±8.35 / 1251.47 ms β 1263.74 / 1276.78 Β±12.75 / 1298.87 ms β no change β β Q11 β 425.69 / 431.41 Β±3.38 / 436.28 ms β 454.23 / 457.47 Β±3.83 / 464.54 ms β 1.06x slower β βββββββββ΄βββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββββ ββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ β Benchmark Summary β β β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββ© β Total Time (HEAD) β 5184.10ms β β Total Time (spill-dedup-view-arrays) β 5404.49ms β β Average Time (HEAD) β 471.28ms β β Average Time (spill-dedup-view-arrays) β 491.32ms β β Queries Faster β 0 β β Queries Slower β 3 β β Queries with No Change β 8 β β Queries with Failure β 0 β ββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ ``` </p> </details> <details><summary>Memory Pool Peaks</summary> Peak `MemoryPool` reservation per query β what DataFusion's accounting believes it reserved. Recorded only when the benchmark runs with `DATAFUSION_RUNTIME_MEMORY_LIMIT` set. Base: `a9b61ab (merge-base)` | Changed: `spill-dedup-view-arrays` **`sort_tpch` β `sort_tpch1`** | Query | Base | Changed | Change | | --- | --- | --- | --- | | 1 | 174.9 MiB | 175.5 MiB | +0.3% | | 2 | 219.9 MiB | 220.6 MiB | +0.3% | | 3 | 505.0 MiB | 503.0 MiB | -0.4% | | 4 | 266.1 MiB | 264.6 MiB | -0.6% | | 5 | 295.0 MiB | 295.0 MiB | +0.0% | | 6 | 356.1 MiB | 355.7 MiB | -0.1% | | 7 | 509.0 MiB | 506.4 MiB | -0.5% | | 8 | 505.8 MiB | 504.1 MiB | -0.3% | | 9 | 505.1 MiB | 505.1 MiB | +0.0% | | 10 | 510.1 MiB | 510.1 MiB | +0.0% | | 11 | 508.9 MiB | 508.9 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 | | --- | --- | --- | --- | --- | --- | | sort_tpch | base (`a9b61ab (merge-base)`) | 510.1 MiB | 1.8 GiB | 1.3 GiB | 3.7Γ | | sort_tpch | changed (`spill-dedup-view-arrays`) | 510.1 MiB | 1.5 GiB | 990.1 MiB | 2.9Γ | </details> <details><summary>Resource Usage</summary> **sort_tpch β base (merge-base)** | Metric | Value | |--------|-------| | Wall time | 30.0s | | Peak memory | 1.8 GiB | | Avg memory | 987.4 MiB | | CPU user | 79.9s | | CPU sys | 16.6s | | Peak spill | 0 B | **sort_tpch β branch** | Metric | Value | |--------|-------| | Wall time | 30.0s | | Peak memory | 1.5 GiB | | Avg memory | 1022.6 MiB | | CPU user | 83.0s | | CPU sys | 17.0s | | 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]
