adriangbot commented on PR #27500: URL: https://github.com/apache/datafusion/pull/27500#issuecomment-6951579153
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/27500#issuecomment-6950997044) **Instance:** `c4-highmem-16` (12 vCPU / 65 GiB) Comparing b27c7e1e6ca28310f5d961ea550fc2f47fd0199c (b27c7e1e6ca28310f5d961ea550fc2f47fd0199c) to b27c7e1e6ca28310f5d961ea550fc2f47fd0199c [diff](https://github.com/apache/datafusion/compare/b27c7e1e6ca28310f5d961ea550fc2f47fd0199c..b27c7e1e6ca28310f5d961ea550fc2f47fd0199c) <details><summary>Run configuration</summary> ```yaml run benchmark clickbench_partitioned env: DATAFUSION_EXECUTION_PARQUET_ENABLE_RLE_TO_DICTIONARY: "true" DATAFUSION_RUNTIME_MEMORY_LIMIT: "32G" baseline: ref: "b27c7e1e6ca28310f5d961ea550fc2f47fd0199c" changed: ref: "b27c7e1e6ca28310f5d961ea550fc2f47fd0199c" env: DATAFUSION_EXECUTION_PARQUET_DICTIONARY_READ_MAX_DISTINCT_VALUES: "5" resources: cpu: "12" memory: "65Gi" arch: "amd64" ``` </details> <details><summary>CPU Details (lscpu)</summary> ``` Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Address sizes: 52 bits physical, 57 bits virtual Byte Order: Little Endian CPU(s): 16 On-line CPU(s) list: 0-15 Vendor ID: GenuineIntel Model name: INTEL(R) XEON(R) PLATINUM 8581C CPU @ 2.30GHz CPU family: 6 Model: 207 Thread(s) per core: 2 Core(s) per socket: 8 Socket(s): 1 Stepping: 2 BogoMIPS: 4600.00 Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid tsc_known_freq pni pclmulqdq monitor ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch ssbd ibrs ibpb stibp ibrs_enhanced fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves avx_vnni avx512_bf16 wbnoinvd arat avx512vbmi umip avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq rdpid cldemote movdiri movdir64b fsrm md_clear serialize tsxldtrk amx_bf16 avx512_fp16 amx_tile amx_int8 arch_capabilities Hypervisor vendor: KVM Virtualization type: full L1d cache: 384 KiB (8 instances) L1i cache: 256 KiB (8 instances) L2 cache: 16 MiB (8 instances) L3 cache: 260 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; usercopy/swapgs barriers and __user pointer sanitization Vulnerability Spectre v2: Mitigation; Enhanced / Automatic IBRS; IBPB conditional; PBRSB-eIBRS SW sequence; BHI BHI_DIS_S 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 rich-T-kid_parquet-dictionary-ndv-threshold -------------------- Benchmark clickbench_partitioned.json -------------------- βββββββββββββ³ββββββββββββββ³ββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ β Query β HEAD β rich-T-kid_parquet-dictionary-ndv-threshold β Change β β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β QQuery 0 β 1.25 ms β 1.57 ms β 1.26x slower β β QQuery 1 β 18.74 ms β 19.47 ms β no change β β QQuery 2 β 52.64 ms β 54.50 ms β no change β β QQuery 3 β 55.73 ms β 56.12 ms β no change β β QQuery 4 β 381.56 ms β 380.45 ms β no change β β QQuery 5 β 1031.92 ms β 441.01 ms β +2.34x faster β β QQuery 6 β 1.29 ms β 1.63 ms β 1.26x slower β β QQuery 7 β 19.87 ms β 20.20 ms β no change β β QQuery 8 β 554.05 ms β 558.42 ms β no change β β QQuery 9 β 659.97 ms β 661.32 ms β no change β β QQuery 10 β 294.95 ms β 112.81 ms β +2.61x faster β β QQuery 11 β 300.86 ms β 130.49 ms β +2.31x faster β β QQuery 12 β 1001.36 ms β 419.34 ms β +2.39x faster β β QQuery 13 β 1520.74 ms β 605.06 ms β +2.51x faster β β QQuery 14 β 1194.33 ms β 420.75 ms β +2.84x faster β β QQuery 15 β 450.41 ms β 445.14 ms β no change β β QQuery 16 β 1890.95 ms β 1046.65 ms β +1.81x faster β β QQuery 17 β 1868.00 ms β 1056.47 ms β +1.77x faster β β QQuery 18 β 3013.92 ms β 2081.49 ms β +1.45x faster β β QQuery 19 β 51.67 ms β 53.01 ms β no change β β QQuery 20 β 1575.65 ms β 805.65 ms β +1.96x faster β β QQuery 21 β 2086.71 ms β 833.72 ms β +2.50x faster β β QQuery 22 β 5119.23 ms β 1608.21 ms β +3.18x faster β β QQuery 23 β 22697.61 ms β 5381.59 ms β +4.22x faster β β QQuery 24 β 143.46 ms β 58.87 ms β +2.44x faster β β QQuery 25 β 536.22 ms β 157.97 ms β +3.39x faster β β QQuery 26 β 143.16 ms β 58.58 ms β +2.44x faster β β QQuery 27 β 1706.80 ms β 812.21 ms β +2.10x faster β β QQuery 28 β 3378.89 ms β 3605.27 ms β 1.07x slower β β QQuery 29 β 51.10 ms β 52.20 ms β no change β β QQuery 30 β 830.43 ms β 469.95 ms β +1.77x faster β β QQuery 31 β 879.39 ms β 499.47 ms β +1.76x faster β β QQuery 32 β 1754.58 ms β 1637.09 ms β +1.07x faster β β QQuery 33 β 4317.18 ms β 2289.63 ms β +1.89x faster β β QQuery 34 β 4395.88 ms β 2275.69 ms β +1.93x faster β β QQuery 35 β 383.59 ms β 378.07 ms β no change β β QQuery 36 β 155.36 ms β 87.50 ms β +1.78x faster β β QQuery 37 β 125.81 ms β 43.64 ms β +2.88x faster β β QQuery 38 β 101.20 ms β 53.76 ms β +1.88x faster β β QQuery 39 β 367.15 ms β 204.39 ms β +1.80x faster β β QQuery 40 β 18.73 ms β 18.41 ms β no change β β QQuery 41 β 16.96 ms β 16.17 ms β no change β β QQuery 42 β 12.81 ms β 13.04 ms β no change β βββββββββββββ΄ββββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ β Benchmark Summary β β β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β Total Time (HEAD) β 65162.07ms β β Total Time (rich-T-kid_parquet-dictionary-ndv-threshold) β 29926.97ms β β Average Time (HEAD) β 1515.40ms β β Average Time (rich-T-kid_parquet-dictionary-ndv-threshold) β 695.98ms β β Queries Faster β 26 β β Queries Slower β 3 β β Queries with No Change β 14 β β Queries with Failure β 0 β ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄βββββββββββββ Distribution per query (min / mean Β±stddev / max): Comparing HEAD and rich-T-kid_parquet-dictionary-ndv-threshold -------------------- Benchmark clickbench_partitioned.json -------------------- βββββββββββββ³ββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ β Query β HEAD β rich-T-kid_parquet-dictionary-ndv-threshold β Change β β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β QQuery 0 β 1.25 / 2.29 Β±1.95 / 6.19 ms β 1.57 / 2.67 Β±2.08 / 6.83 ms β 1.17x slower β β QQuery 1 β 18.74 / 19.90 Β±1.14 / 21.82 ms β 19.47 / 20.30 Β±0.79 / 21.76 ms β no change β β QQuery 2 β 52.64 / 56.55 Β±3.32 / 62.18 ms β 54.50 / 57.48 Β±3.03 / 62.12 ms β no change β β QQuery 3 β 55.73 / 56.78 Β±1.20 / 58.47 ms β 56.12 / 56.99 Β±1.00 / 58.83 ms β no change β β QQuery 4 β 381.56 / 392.58 Β±8.42 / 401.05 ms β 380.45 / 398.65 Β±13.70 / 417.17 ms β no change β β QQuery 5 β 1031.92 / 1039.64 Β±10.73 / 1060.90 ms β 441.01 / 451.69 Β±6.41 / 459.01 ms β +2.30x faster β β QQuery 6 β 1.29 / 1.58 Β±0.38 / 2.28 ms β 1.63 / 1.89 Β±0.34 / 2.53 ms β 1.20x slower β β QQuery 7 β 19.87 / 20.19 Β±0.30 / 20.73 ms β 20.20 / 20.91 Β±0.49 / 21.71 ms β no change β β QQuery 8 β 554.05 / 565.89 Β±7.53 / 577.77 ms β 558.42 / 570.92 Β±11.69 / 589.65 ms β no change β β QQuery 9 β 659.97 / 686.60 Β±25.27 / 728.09 ms β 661.32 / 688.62 Β±21.30 / 707.78 ms β no change β β QQuery 10 β 294.95 / 298.88 Β±3.17 / 302.83 ms β 112.81 / 116.07 Β±3.22 / 120.95 ms β +2.58x faster β β QQuery 11 β 300.86 / 314.07 Β±11.95 / 329.87 ms β 130.49 / 133.48 Β±3.18 / 139.64 ms β +2.35x faster β β QQuery 12 β 1001.36 / 1024.10 Β±19.25 / 1056.30 ms β 419.34 / 443.63 Β±14.16 / 457.60 ms β +2.31x faster β β QQuery 13 β 1520.74 / 1532.17 Β±11.48 / 1547.27 ms β 605.06 / 628.17 Β±18.87 / 650.64 ms β +2.44x faster β β QQuery 14 β 1194.33 / 1216.00 Β±16.59 / 1238.98 ms β 420.75 / 449.36 Β±26.05 / 488.83 ms β +2.71x faster β β QQuery 15 β 450.41 / 467.07 Β±15.29 / 492.93 ms β 445.14 / 453.81 Β±8.88 / 470.08 ms β no change β β QQuery 16 β 1890.95 / 1946.85 Β±29.61 / 1976.73 ms β 1046.65 / 1073.58 Β±15.93 / 1089.13 ms β +1.81x faster β β QQuery 17 β 1868.00 / 1894.02 Β±13.46 / 1906.04 ms β 1056.47 / 1070.97 Β±13.06 / 1092.41 ms β +1.77x faster β β QQuery 18 β 3013.92 / 3076.66 Β±38.31 / 3134.09 ms β 2081.49 / 2103.79 Β±16.70 / 2122.52 ms β +1.46x faster β β QQuery 19 β 51.67 / 54.09 Β±3.03 / 59.95 ms β 53.01 / 57.86 Β±8.49 / 74.83 ms β 1.07x slower β β QQuery 20 β 1575.65 / 1605.85 Β±23.55 / 1645.75 ms β 805.65 / 828.19 Β±15.89 / 855.07 ms β +1.94x faster β β QQuery 21 β 2086.71 / 2110.08 Β±17.41 / 2135.60 ms β 833.72 / 853.08 Β±10.90 / 866.33 ms β +2.47x faster β β QQuery 22 β 5119.23 / 5165.19 Β±38.77 / 5230.97 ms β 1608.21 / 1634.94 Β±17.58 / 1663.21 ms β +3.16x faster β β QQuery 23 β 22697.61 / 22973.65 Β±167.35 / 23193.37 ms β 5381.59 / 5433.39 Β±39.77 / 5496.62 ms β +4.23x faster β β QQuery 24 β 143.46 / 161.17 Β±24.60 / 207.62 ms β 58.87 / 60.96 Β±3.64 / 68.22 ms β +2.64x faster β β QQuery 25 β 536.22 / 544.44 Β±6.06 / 554.61 ms β 157.97 / 162.14 Β±3.05 / 166.62 ms β +3.36x faster β β QQuery 26 β 143.16 / 145.41 Β±2.33 / 149.36 ms β 58.58 / 62.14 Β±2.19 / 64.61 ms β +2.34x faster β β QQuery 27 β 1706.80 / 1724.02 Β±11.30 / 1736.78 ms β 812.21 / 822.08 Β±10.41 / 840.50 ms β +2.10x faster β β QQuery 28 β 3378.89 / 3430.60 Β±31.75 / 3474.11 ms β 3605.27 / 3634.83 Β±29.90 / 3677.48 ms β 1.06x slower β β QQuery 29 β 51.10 / 64.07 Β±11.91 / 84.70 ms β 52.20 / 61.93 Β±15.49 / 92.73 ms β no change β β QQuery 30 β 830.43 / 843.86 Β±11.24 / 863.76 ms β 469.95 / 484.94 Β±11.55 / 505.37 ms β +1.74x faster β β QQuery 31 β 879.39 / 891.02 Β±11.27 / 907.76 ms β 499.47 / 511.91 Β±18.31 / 548.32 ms β +1.74x faster β β QQuery 32 β 1754.58 / 1812.67 Β±73.87 / 1956.10 ms β 1637.09 / 1686.84 Β±34.91 / 1737.14 ms β +1.07x faster β β QQuery 33 β 4317.18 / 4416.00 Β±73.63 / 4528.98 ms β 2289.63 / 2312.74 Β±15.70 / 2330.91 ms β +1.91x faster β β QQuery 34 β 4395.88 / 4481.34 Β±57.32 / 4571.80 ms β 2275.69 / 2324.54 Β±32.32 / 2375.50 ms β +1.93x faster β β QQuery 35 β 383.59 / 399.10 Β±16.48 / 430.03 ms β 378.07 / 391.39 Β±11.50 / 405.86 ms β no change β β QQuery 36 β 155.36 / 166.17 Β±7.35 / 174.61 ms β 87.50 / 97.78 Β±8.41 / 105.85 ms β +1.70x faster β β QQuery 37 β 125.81 / 131.10 Β±3.60 / 134.31 ms β 43.64 / 49.75 Β±8.37 / 65.89 ms β +2.64x faster β β QQuery 38 β 101.20 / 107.53 Β±7.19 / 121.09 ms β 53.76 / 59.22 Β±3.19 / 63.58 ms β +1.82x faster β β QQuery 39 β 367.15 / 383.01 Β±11.26 / 395.09 ms β 204.39 / 211.10 Β±7.87 / 225.47 ms β +1.81x faster β β QQuery 40 β 18.73 / 23.09 Β±3.41 / 28.10 ms β 18.41 / 20.86 Β±1.81 / 23.04 ms β +1.11x faster β β QQuery 41 β 16.96 / 17.51 Β±0.46 / 18.30 ms β 16.17 / 17.44 Β±1.32 / 19.76 ms β no change β β QQuery 42 β 12.81 / 13.68 Β±1.09 / 15.80 ms β 13.04 / 13.63 Β±0.97 / 15.57 ms β no change β βββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³βββββββββββββ β Benchmark Summary β β β‘ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β Total Time (HEAD) β 66276.48ms β β Total Time (rich-T-kid_parquet-dictionary-ndv-threshold) β 30536.66ms β β Average Time (HEAD) β 1541.31ms β β Average Time (rich-T-kid_parquet-dictionary-ndv-threshold) β 710.15ms β β Queries Faster β 27 β β Queries Slower β 4 β β Queries with No Change β 12 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: `b27c7e1e6ca28310f5d961ea550fc2f47fd0199c` | Changed: `b27c7e1e6ca28310f5d961ea550fc2f47fd0199c` **`clickbench_partitioned`** | Query | Base | Changed | Change | | --- | --- | --- | --- | | Query 0 | 0 B | 0 B | 0.0% | | Query 1 | 104 B | 104 B | +0.0% | | Query 2 | 936 B | 936 B | +0.0% | | Query 3 | 312 B | 312 B | +0.0% | | Query 4 | 830.0 MiB | 831.6 MiB | +0.2% | | Query 5 | 1.4 GiB | 1.2 GiB | -12.3% | | Query 6 | 0 B | 0 B | 0.0% | | Query 7 | 60.0 MiB | 40.4 MiB | -32.7% | | Query 8 | 921.3 MiB | 930.0 MiB | +0.9% | | Query 9 | 677.7 MiB | 677.5 MiB | -0.0% | | Query 10 | 72.7 MiB | 107.7 MiB | +48.1% | | Query 11 | 78.3 MiB | 115.7 MiB | +47.8% | | Query 12 | 1.5 GiB | 1.4 GiB | -5.5% | | Query 13 | 1.5 GiB | 1.5 GiB | -0.3% | | Query 14 | 1.6 GiB | 1.3 GiB | -17.1% | | Query 15 | 1.1 GiB | 1.1 GiB | -1.4% | | Query 16 | 3.1 GiB | 3.4 GiB | +10.6% | | Query 17 | 3.3 GiB | 3.3 GiB | +1.6% | | Query 18 | 6.9 GiB | 7.3 GiB | +5.5% | | Query 19 | 0 B | 0 B | 0.0% | | Query 20 | 104 B | 104 B | +0.0% | | Query 21 | 1.5 MiB | 4.3 MiB | +185.4% | | Query 22 | 2.2 MiB | 7.7 MiB | +244.6% | | Query 23 | 12.9 MiB | 28.1 MiB | +119.0% | | Query 24 | 6.0 MiB | 60.1 MiB | +901.9% | | Query 25 | 7.1 MiB | 180.6 MiB | +2451.5% | | Query 26 | 7.3 MiB | 64.6 MiB | +789.9% | | Query 27 | 3.0 MiB | 2.8 MiB | -5.1% | | Query 28 | 2.0 GiB | 1.7 GiB | -16.5% | | Query 29 | 624 B | 624 B | +0.0% | | Query 30 | 648.5 MiB | 718.7 MiB | +10.8% | | Query 31 | 1.5 GiB | 1.5 GiB | +0.0% | | Query 32 | 7.1 GiB | 7.1 GiB | -0.3% | | Query 33 | 9.6 GiB | 8.0 GiB | -16.7% | | Query 34 | 9.8 GiB | 8.0 GiB | -19.0% | | Query 35 | 664.7 MiB | 652.9 MiB | -1.8% | | Query 36 | 99.0 MiB | 101.6 MiB | +2.6% | | Query 37 | 7.6 MiB | 7.9 MiB | +3.6% | | Query 38 | 6.1 MiB | 5.5 MiB | -11.2% | | Query 39 | 212.9 MiB | 302.0 MiB | +41.9% | | Query 40 | 2.3 MiB | 2.1 MiB | -9.5% | | Query 41 | 3.3 MiB | 3.3 MiB | +0.0% | | Query 42 | 1.8 MiB | 1.6 MiB | -10.8% | **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 | | --- | --- | --- | --- | --- | --- | | clickbench_partitioned | base (`b27c7e1e6ca28310f5d961ea550fc2f47fd0199c`) | 9.8 GiB | 12.8 GiB | 2.9 GiB | 1.3Γ | | clickbench_partitioned | changed (`b27c7e1e6ca28310f5d961ea550fc2f47fd0199c`) | 8.0 GiB | 10.4 GiB | 2.3 GiB | 1.3Γ | </details> <details><summary>Resource Usage</summary> **clickbench_partitioned β base (merge-base)** | Metric | Value | |--------|-------| | Wall time | 335.1s | | Peak memory | 12.8 GiB | | Avg memory | 5.4 GiB | | CPU user | 3331.5s | | CPU sys | 298.5s | | Peak spill | 0 B | **clickbench_partitioned β branch** | Metric | Value | |--------|-------| | Wall time | 155.0s | | Peak memory | 10.4 GiB | | Avg memory | 4.3 GiB | | CPU user | 1533.2s | | CPU sys | 175.9s | | 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]
