adriangbot commented on PR #27500: URL: https://github.com/apache/datafusion/pull/27500#issuecomment-6939695566
π€ Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/27500#issuecomment-6939383888) **Instance:** `c4-highmem-16` (12 vCPU / 65 GiB) Comparing 432586c6f82b2a7f96933dc0501cf02f2a2a1eca (432586c6f82b2a7f96933dc0501cf02f2a2a1eca) to 432586c6f82b2a7f96933dc0501cf02f2a2a1eca [diff](https://github.com/apache/datafusion/compare/432586c6f82b2a7f96933dc0501cf02f2a2a1eca..432586c6f82b2a7f96933dc0501cf02f2a2a1eca) <details><summary>Run configuration</summary> ```yaml run benchmark tpch env: DATAFUSION_EXECUTION_PARQUET_ENABLE_RLE_TO_DICTIONARY: "true" DATAFUSION_RUNTIME_MEMORY_LIMIT: "32G" baseline: ref: "432586c6f82b2a7f96933dc0501cf02f2a2a1eca" changed: ref: "432586c6f82b2a7f96933dc0501cf02f2a2a1eca" 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 tpch_sf1.json -------------------- βββββββββββββ³βββββββββββ³ββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ β Query β HEAD β rich-T-kid_parquet-dictionary-ndv-threshold β Change β β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β QQuery 1 β 69.34 ms β 75.51 ms β 1.09x slower β β QQuery 2 β 20.87 ms β 20.67 ms β no change β β QQuery 3 β 41.80 ms β 42.03 ms β no change β β QQuery 4 β 25.01 ms β 25.14 ms β no change β β QQuery 5 β 51.00 ms β 51.90 ms β no change β β QQuery 6 β 20.53 ms β 21.78 ms β 1.06x slower β β QQuery 7 β 63.53 ms β 59.23 ms β +1.07x faster β β QQuery 8 β 55.87 ms β 57.62 ms β no change β β QQuery 9 β 78.60 ms β 77.95 ms β no change β β QQuery 10 β 75.71 ms β 57.68 ms β +1.31x faster β β QQuery 11 β 15.49 ms β 15.45 ms β no change β β QQuery 12 β 29.49 ms β 28.17 ms β no change β β QQuery 13 β 58.35 ms β 47.85 ms β +1.22x faster β β QQuery 14 β 35.42 ms β 33.44 ms β +1.06x faster β β QQuery 15 β 45.72 ms β 48.74 ms β 1.07x slower β β QQuery 16 β 15.52 ms β 15.98 ms β no change β β QQuery 17 β 89.38 ms β 92.14 ms β no change β β QQuery 18 β 84.61 ms β 61.03 ms β +1.39x faster β β QQuery 19 β 45.14 ms β 51.07 ms β 1.13x slower β β QQuery 20 β 47.83 ms β 44.09 ms β +1.08x faster β β QQuery 21 β 83.56 ms β 87.60 ms β no change β β QQuery 22 β 17.97 ms β 17.15 ms β no change β βββββββββββββ΄βββββββββββ΄ββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ β Benchmark Summary β β β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β Total Time (HEAD) β 1070.74ms β β Total Time (rich-T-kid_parquet-dictionary-ndv-threshold) β 1032.22ms β β Average Time (HEAD) β 48.67ms β β Average Time (rich-T-kid_parquet-dictionary-ndv-threshold) β 46.92ms β β Queries Faster β 6 β β Queries Slower β 4 β β Queries with No Change β 12 β β Queries with Failure β 0 β ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ Distribution per query (min / mean Β±stddev / max): Comparing HEAD and rich-T-kid_parquet-dictionary-ndv-threshold -------------------- Benchmark tpch_sf1.json -------------------- βββββββββββββ³βββββββββββββββββββββββββββββββββ³ββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββββββ β Query β HEAD β rich-T-kid_parquet-dictionary-ndv-threshold β Change β β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β QQuery 1 β 69.34 / 76.89 Β±5.77 / 84.34 ms β 75.51 / 79.15 Β±2.32 / 82.31 ms β no change β β QQuery 2 β 20.87 / 22.58 Β±1.95 / 26.32 ms β 20.67 / 22.57 Β±1.88 / 26.15 ms β no change β β QQuery 3 β 41.80 / 43.40 Β±0.84 / 44.27 ms β 42.03 / 43.94 Β±1.47 / 46.41 ms β no change β β QQuery 4 β 25.01 / 25.69 Β±0.60 / 26.65 ms β 25.14 / 25.75 Β±0.58 / 26.68 ms β no change β β QQuery 5 β 51.00 / 52.91 Β±1.55 / 54.98 ms β 51.90 / 52.91 Β±0.79 / 54.22 ms β no change β β QQuery 6 β 20.53 / 22.42 Β±1.16 / 23.81 ms β 21.78 / 23.02 Β±1.03 / 24.85 ms β no change β β QQuery 7 β 63.53 / 64.50 Β±0.84 / 65.97 ms β 59.23 / 62.74 Β±2.13 / 65.50 ms β no change β β QQuery 8 β 55.87 / 59.60 Β±2.57 / 63.55 ms β 57.62 / 59.94 Β±1.48 / 61.54 ms β no change β β QQuery 9 β 78.60 / 80.63 Β±1.13 / 81.93 ms β 77.95 / 79.46 Β±1.71 / 82.25 ms β no change β β QQuery 10 β 75.71 / 81.31 Β±3.40 / 85.82 ms β 57.68 / 59.98 Β±1.86 / 63.14 ms β +1.36x faster β β QQuery 11 β 15.49 / 16.64 Β±0.91 / 17.92 ms β 15.45 / 16.50 Β±0.96 / 17.97 ms β no change β β QQuery 12 β 29.49 / 30.27 Β±0.71 / 31.59 ms β 28.17 / 28.85 Β±0.69 / 30.05 ms β no change β β QQuery 13 β 58.35 / 62.95 Β±6.21 / 75.21 ms β 47.85 / 49.45 Β±2.22 / 53.80 ms β +1.27x faster β β QQuery 14 β 35.42 / 36.17 Β±0.93 / 37.81 ms β 33.44 / 36.17 Β±2.02 / 39.65 ms β no change β β QQuery 15 β 45.72 / 47.78 Β±1.51 / 50.23 ms β 48.74 / 49.88 Β±0.85 / 51.29 ms β no change β β QQuery 16 β 15.52 / 15.87 Β±0.22 / 16.15 ms β 15.98 / 16.73 Β±0.65 / 17.70 ms β 1.05x slower β β QQuery 17 β 89.38 / 93.12 Β±2.12 / 95.95 ms β 92.14 / 93.62 Β±0.92 / 94.40 ms β no change β β QQuery 18 β 84.61 / 90.42 Β±4.99 / 96.41 ms β 61.03 / 64.60 Β±2.04 / 67.28 ms β +1.40x faster β β QQuery 19 β 45.14 / 46.91 Β±1.04 / 47.94 ms β 51.07 / 51.86 Β±0.64 / 52.75 ms β 1.11x slower β β QQuery 20 β 47.83 / 50.50 Β±3.45 / 57.24 ms β 44.09 / 46.26 Β±1.66 / 48.99 ms β +1.09x faster β β QQuery 21 β 83.56 / 85.67 Β±2.51 / 90.31 ms β 87.60 / 89.98 Β±3.55 / 97.00 ms β 1.05x slower β β QQuery 22 β 17.97 / 19.72 Β±1.02 / 20.85 ms β 17.15 / 17.65 Β±0.39 / 18.26 ms β +1.12x faster β βββββββββββββ΄βββββββββββββββββββββββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββββββ ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ³ββββββββββββ β Benchmark Summary β β β‘βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ© β Total Time (HEAD) β 1125.95ms β β Total Time (rich-T-kid_parquet-dictionary-ndv-threshold) β 1071.03ms β β Average Time (HEAD) β 51.18ms β β Average Time (rich-T-kid_parquet-dictionary-ndv-threshold) β 48.68ms β β Queries Faster β 5 β β Queries Slower β 3 β β Queries with No Change β 14 β β Queries with Failure β 0 β ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ΄ββββββββββββ ``` </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: `432586c6f82b2a7f96933dc0501cf02f2a2a1eca` | Changed: `432586c6f82b2a7f96933dc0501cf02f2a2a1eca` **`tpch` β `tpch_sf1`** | Query | Base | Changed | Change | | --- | --- | --- | --- | | Query 1 | 21.2 MiB | 31.4 MiB | +47.7% | | Query 2 | 23.1 MiB | 15.9 MiB | -31.3% | | Query 3 | 11.1 MiB | 11.2 MiB | +0.2% | | Query 4 | 22.2 MiB | 33.4 MiB | +50.3% | | Query 5 | 42.7 MiB | 43.2 MiB | +1.0% | | Query 6 | 832 B | 832 B | +0.0% | | Query 7 | 103.6 MiB | 107.1 MiB | +3.4% | | Query 8 | 33.1 MiB | 39.8 MiB | +20.2% | | Query 9 | 140.9 MiB | 140.6 MiB | -0.2% | | Query 10 | 50.0 MiB | 43.9 MiB | -12.2% | | Query 11 | 31.7 MiB | 51.2 MiB | +61.4% | | Query 12 | 22.5 MiB | 22.6 MiB | +0.5% | | Query 13 | 92.3 MiB | 72.3 MiB | -21.6% | | Query 14 | 6.7 MiB | 6.7 MiB | +0.0% | | Query 15 | 13.0 MiB | 12.4 MiB | -4.5% | | Query 16 | 42.5 MiB | 45.0 MiB | +6.0% | | Query 17 | 153.3 MiB | 154.5 MiB | +0.8% | | Query 18 | 116.3 MiB | 167.1 MiB | +43.6% | | Query 19 | 209.7 KiB | 725.7 KiB | +246.1% | | Query 20 | 59.8 MiB | 57.3 MiB | -4.2% | | Query 21 | 36.1 MiB | 42.4 MiB | +17.6% | | Query 22 | 31.5 MiB | 32.2 MiB | +2.1% | **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 (`432586c6f82b2a7f96933dc0501cf02f2a2a1eca`) | 153.3 MiB | 974.0 MiB | 820.7 MiB | 6.4Γ | | tpch | changed (`432586c6f82b2a7f96933dc0501cf02f2a2a1eca`) | 167.1 MiB | 1.1 GiB | 991.0 MiB | 6.9Γ | </details> <details><summary>Resource Usage</summary> **tpch β base (merge-base)** | Metric | Value | |--------|-------| | Wall time | 10.0s | | Peak memory | 974.0 MiB | | Avg memory | 391.7 MiB | | CPU user | 35.2s | | CPU sys | 1.9s | | Peak spill | 0 B | **tpch β branch** | Metric | Value | |--------|-------| | Wall time | 10.0s | | Peak memory | 1.1 GiB | | Avg memory | 424.6 MiB | | CPU user | 34.6s | | CPU sys | 2.0s | | 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]
