You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@datafusion.apache.org by "Dandandan (via GitHub)" <gi...@apache.org> on 2026/05/01 18:35:15 UTC

[PR] perf: speed up primitive group value interning [datafusion]

Dandandan opened a new pull request, #21977:
URL: https://github.com/apache/datafusion/pull/21977

   ## Which issue does this PR close?
   
   N/A.
   
   ## Rationale for this change
   
   `GroupValuesPrimitive` stored hashes in the hash table. For single primitive group keys, storing the primitive value in the table lets hashbrown recompute hashes from the value and removes separate hash storage and values-vector lookups in the hot interning path.
   
   ## What changes are included in this PR?
   
   - Store `(group_index, value)` in `GroupValuesPrimitive` instead of `(group_index, hash)` plus a separate values vector.
   - Reconstruct emitted primitive arrays from hash table entries in group-index order.
   - Add focused `GroupValuesPrimitive_intern` benchmark cases and emit/reindex regression tests.
   
   ## Are these changes tested?
   
   Yes:
   - `cargo fmt --all`
   - `ci/scripts/rust_clippy.sh`
   - `cargo test -p datafusion-physical-plan --lib group_values`
   - `cargo test -p datafusion-physical-plan --lib primitive_emit`
   - `cargo clippy -p datafusion-physical-plan --features test_utils --bench aggregate_vectorized -- -D warnings`
   - `cargo clippy -p datafusion-physical-plan --lib --tests --features test_utils -- -D warnings`
   
   Benchmarks vs original stored-hash baseline:
   - `GroupValuesPrimitive_intern/low_cardinality`: 138.57 us -> 117.75 us (-15.48%)
   - `GroupValuesPrimitive_intern/high_cardinality`: 1.1006 ms -> 1.0879 ms (no statistically significant change)
   - `aggregate_query_group_by_u64 15 12`: 785.50 us -> 741.78 us (-5.06%)
   
   Note: `cargo clippy --all-targets --all-features -- -D warnings` currently fails in this workspace because `--all-features` enables both benchmark allocator features `snmalloc` and `mimalloc`; `ci/scripts/rust_clippy.sh` is the repository CI clippy command and passed.
   
   ## Are there any user-facing changes?
   
   No.


-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4365794880

   run benchmark clickbench_extended tpch10


-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363658396

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4363624010)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark tpcds_sf1.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                     HEAD โ”ƒ       speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 1  โ”‚              7.21 / 7.79 ยฑ0.79 / 9.35 ms โ”‚              7.25 / 7.77 ยฑ0.73 / 9.17 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚        145.71 / 146.95 ยฑ0.77 / 147.77 ms โ”‚        143.09 / 143.44 ยฑ0.33 / 144.00 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚        114.05 / 114.90 ยฑ0.64 / 116.00 ms โ”‚        113.08 / 114.40 ยฑ0.95 / 115.85 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚    1344.71 / 1364.48 ยฑ13.82 / 1380.20 ms โ”‚    1373.26 / 1382.47 ยฑ10.09 / 1394.92 ms โ”‚     no change โ”‚
   โ”‚ QQuery 5  โ”‚        176.30 / 177.93 ยฑ0.98 / 178.90 ms โ”‚        173.51 / 175.14 ยฑ1.50 / 177.90 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚        146.60 / 150.32 ยฑ3.00 / 153.97 ms โ”‚        153.20 / 156.45 ยฑ2.24 / 159.60 ms โ”‚     no change โ”‚
   โ”‚ QQuery 7  โ”‚        328.75 / 330.34 ยฑ1.30 / 332.01 ms โ”‚        324.86 / 326.23 ยฑ0.95 / 327.43 ms โ”‚     no change โ”‚
   โ”‚ QQuery 8  โ”‚        115.18 / 115.71 ยฑ0.54 / 116.70 ms โ”‚        114.28 / 114.76 ยฑ0.62 / 115.97 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚          93.65 / 99.44 ยฑ3.47 / 104.31 ms โ”‚          96.36 / 99.84 ยฑ2.95 / 103.68 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚        104.44 / 106.26 ยฑ1.47 / 108.75 ms โ”‚        101.86 / 103.25 ยฑ1.17 / 105.31 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚        951.88 / 965.10 ยฑ7.98 / 973.81 ms โ”‚        965.72 / 979.90 ยฑ8.26 / 988.79 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚           44.25 / 45.02 ยฑ0.54 / 45.65 ms โ”‚           44.25 / 45.63 ยฑ1.67 / 48.89 ms โ”‚     no change โ”‚
   โ”‚ QQuery 13 โ”‚        398.00 / 399.01 ยฑ1.17 / 401.31 ms โ”‚        396.08 / 398.97 ยฑ2.74 / 403.95 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚     1043.54 / 1048.05 ยฑ3.42 / 1053.90 ms โ”‚     1039.76 / 1044.35 ยฑ3.83 / 1048.96 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚           15.28 / 15.55 ยฑ0.38 / 16.28 ms โ”‚           14.58 / 15.00 ยฑ0.31 / 15.37 ms โ”‚     no change โ”‚
   โ”‚ QQuery 16 โ”‚              7.62 / 7.82 ยฑ0.13 / 8.01 ms โ”‚              7.32 / 7.57 ยฑ0.23 / 8.01 ms โ”‚     no change โ”‚
   โ”‚ QQuery 17 โ”‚        208.36 / 209.88 ยฑ1.39 / 211.67 ms โ”‚        204.04 / 207.37 ยฑ3.06 / 212.53 ms โ”‚     no change โ”‚
   โ”‚ QQuery 18 โ”‚        125.83 / 126.74 ยฑ1.00 / 128.52 ms โ”‚        124.77 / 125.71 ยฑ1.41 / 128.50 ms โ”‚     no change โ”‚
   โ”‚ QQuery 19 โ”‚        157.61 / 158.35 ยฑ0.48 / 158.90 ms โ”‚        156.18 / 157.13 ยฑ0.86 / 158.20 ms โ”‚     no change โ”‚
   โ”‚ QQuery 20 โ”‚           13.61 / 13.71 ยฑ0.11 / 13.90 ms โ”‚           13.66 / 14.43 ยฑ1.21 / 16.83 ms โ”‚  1.05x slower โ”‚
   โ”‚ QQuery 21 โ”‚           19.78 / 20.35 ยฑ0.91 / 22.17 ms โ”‚           19.82 / 19.92 ยฑ0.08 / 20.07 ms โ”‚     no change โ”‚
   โ”‚ QQuery 22 โ”‚        480.76 / 483.43 ยฑ2.29 / 486.98 ms โ”‚        485.46 / 489.85 ยฑ3.55 / 494.28 ms โ”‚     no change โ”‚
   โ”‚ QQuery 23 โ”‚     1055.32 / 1062.19 ยฑ5.32 / 1071.36 ms โ”‚     1054.18 / 1067.79 ยฑ9.74 / 1082.34 ms โ”‚     no change โ”‚
   โ”‚ QQuery 24 โ”‚        652.26 / 654.80 ยฑ1.88 / 657.43 ms โ”‚        654.60 / 662.29 ยฑ6.10 / 669.68 ms โ”‚     no change โ”‚
   โ”‚ QQuery 25 โ”‚        310.76 / 313.41 ยฑ1.87 / 315.52 ms โ”‚        312.72 / 315.78 ยฑ2.34 / 318.86 ms โ”‚     no change โ”‚
   โ”‚ QQuery 26 โ”‚           76.86 / 77.46 ยฑ0.53 / 78.44 ms โ”‚           76.93 / 79.15 ยฑ3.34 / 85.73 ms โ”‚     no change โ”‚
   โ”‚ QQuery 27 โ”‚             7.12 / 9.04 ยฑ3.41 / 15.86 ms โ”‚              7.45 / 7.50 ยฑ0.06 / 7.62 ms โ”‚ +1.21x faster โ”‚
   โ”‚ QQuery 28 โ”‚        150.17 / 151.30 ยฑ0.74 / 152.42 ms โ”‚        149.17 / 150.31 ยฑ1.30 / 152.84 ms โ”‚     no change โ”‚
   โ”‚ QQuery 29 โ”‚        250.17 / 253.42 ยฑ2.38 / 257.31 ms โ”‚        252.22 / 258.54 ยฑ7.53 / 269.80 ms โ”‚     no change โ”‚
   โ”‚ QQuery 30 โ”‚           42.97 / 43.41 ยฑ0.52 / 44.23 ms โ”‚           42.29 / 42.94 ยฑ0.45 / 43.49 ms โ”‚     no change โ”‚
   โ”‚ QQuery 31 โ”‚        166.48 / 167.79 ยฑ1.31 / 170.26 ms โ”‚        167.45 / 169.78 ยฑ1.42 / 171.72 ms โ”‚     no change โ”‚
   โ”‚ QQuery 32 โ”‚           13.85 / 14.09 ยฑ0.18 / 14.33 ms โ”‚           13.57 / 13.88 ยฑ0.28 / 14.30 ms โ”‚     no change โ”‚
   โ”‚ QQuery 33 โ”‚        140.06 / 141.26 ยฑ1.02 / 142.53 ms โ”‚        139.28 / 141.45 ยฑ2.33 / 145.95 ms โ”‚     no change โ”‚
   โ”‚ QQuery 34 โ”‚              7.36 / 7.45 ยฑ0.11 / 7.66 ms โ”‚              7.44 / 7.54 ยฑ0.07 / 7.61 ms โ”‚     no change โ”‚
   โ”‚ QQuery 35 โ”‚        102.42 / 104.45 ยฑ2.66 / 109.57 ms โ”‚        101.70 / 103.37 ยฑ1.39 / 105.52 ms โ”‚     no change โ”‚
   โ”‚ QQuery 36 โ”‚              7.22 / 7.35 ยฑ0.09 / 7.49 ms โ”‚              6.95 / 7.10 ยฑ0.15 / 7.31 ms โ”‚     no change โ”‚
   โ”‚ QQuery 37 โ”‚              8.45 / 8.51 ยฑ0.06 / 8.63 ms โ”‚              8.50 / 8.55 ยฑ0.05 / 8.64 ms โ”‚     no change โ”‚
   โ”‚ QQuery 38 โ”‚           86.44 / 87.99 ยฑ0.89 / 89.19 ms โ”‚           87.13 / 90.24 ยฑ3.39 / 96.52 ms โ”‚     no change โ”‚
   โ”‚ QQuery 39 โ”‚        120.28 / 123.14 ยฑ4.35 / 131.78 ms โ”‚        121.07 / 123.84 ยฑ2.09 / 126.49 ms โ”‚     no change โ”‚
   โ”‚ QQuery 40 โ”‚           95.72 / 97.78 ยฑ1.57 / 99.70 ms โ”‚         97.57 / 100.79 ยฑ2.38 / 103.33 ms โ”‚     no change โ”‚
   โ”‚ QQuery 41 โ”‚           15.03 / 15.19 ยฑ0.12 / 15.34 ms โ”‚           15.21 / 15.36 ยฑ0.14 / 15.62 ms โ”‚     no change โ”‚
   โ”‚ QQuery 42 โ”‚        106.71 / 108.17 ยฑ0.91 / 109.51 ms โ”‚        107.98 / 108.44 ยฑ0.37 / 108.93 ms โ”‚     no change โ”‚
   โ”‚ QQuery 43 โ”‚              5.90 / 6.05 ยฑ0.13 / 6.24 ms โ”‚              5.90 / 6.11 ยฑ0.17 / 6.41 ms โ”‚     no change โ”‚
   โ”‚ QQuery 44 โ”‚           11.27 / 11.47 ยฑ0.11 / 11.61 ms โ”‚           11.53 / 11.60 ยฑ0.06 / 11.67 ms โ”‚     no change โ”‚
   โ”‚ QQuery 45 โ”‚           45.33 / 45.64 ยฑ0.21 / 45.90 ms โ”‚           45.95 / 46.79 ยฑ1.42 / 49.61 ms โ”‚     no change โ”‚
   โ”‚ QQuery 46 โ”‚            8.97 / 10.04 ยฑ1.82 / 13.67 ms โ”‚              8.76 / 8.94 ยฑ0.16 / 9.13 ms โ”‚ +1.12x faster โ”‚
   โ”‚ QQuery 47 โ”‚        737.52 / 744.09 ยฑ3.84 / 748.03 ms โ”‚        753.13 / 758.52 ยฑ5.63 / 768.53 ms โ”‚     no change โ”‚
   โ”‚ QQuery 48 โ”‚        276.90 / 285.57 ยฑ6.81 / 296.88 ms โ”‚        275.35 / 279.48 ยฑ4.30 / 287.63 ms โ”‚     no change โ”‚
   โ”‚ QQuery 49 โ”‚        248.17 / 248.86 ยฑ1.04 / 250.90 ms โ”‚        244.37 / 246.25 ยฑ1.58 / 249.18 ms โ”‚     no change โ”‚
   โ”‚ QQuery 50 โ”‚        193.07 / 196.26 ยฑ5.05 / 206.31 ms โ”‚        192.95 / 197.37 ยฑ3.25 / 201.66 ms โ”‚     no change โ”‚
   โ”‚ QQuery 51 โ”‚        172.18 / 174.27 ยฑ1.59 / 176.88 ms โ”‚        172.61 / 176.75 ยฑ3.88 / 184.08 ms โ”‚     no change โ”‚
   โ”‚ QQuery 52 โ”‚        107.95 / 108.59 ยฑ0.51 / 109.25 ms โ”‚        108.09 / 110.00 ยฑ2.23 / 112.82 ms โ”‚     no change โ”‚
   โ”‚ QQuery 53 โ”‚        104.07 / 105.16 ยฑ1.72 / 108.58 ms โ”‚        103.35 / 103.78 ยฑ0.29 / 104.17 ms โ”‚     no change โ”‚
   โ”‚ QQuery 54 โ”‚        148.08 / 149.72 ยฑ2.52 / 154.72 ms โ”‚        144.56 / 148.95 ยฑ3.02 / 153.30 ms โ”‚     no change โ”‚
   โ”‚ QQuery 55 โ”‚        107.31 / 107.87 ยฑ0.36 / 108.33 ms โ”‚        106.46 / 107.28 ยฑ0.44 / 107.76 ms โ”‚     no change โ”‚
   โ”‚ QQuery 56 โ”‚        139.76 / 142.38 ยฑ1.98 / 144.76 ms โ”‚        141.35 / 142.54 ยฑ1.29 / 144.98 ms โ”‚     no change โ”‚
   โ”‚ QQuery 57 โ”‚        168.07 / 170.00 ยฑ2.27 / 174.28 ms โ”‚        167.63 / 169.96 ยฑ2.76 / 175.37 ms โ”‚     no change โ”‚
   โ”‚ QQuery 58 โ”‚        246.49 / 248.49 ยฑ1.44 / 250.92 ms โ”‚        245.57 / 246.71 ยฑ1.20 / 248.85 ms โ”‚     no change โ”‚
   โ”‚ QQuery 59 โ”‚        195.66 / 198.82 ยฑ2.67 / 203.46 ms โ”‚        195.51 / 197.68 ยฑ2.00 / 201.08 ms โ”‚     no change โ”‚
   โ”‚ QQuery 60 โ”‚        142.36 / 143.79 ยฑ1.39 / 145.79 ms โ”‚        140.71 / 142.96 ยฑ1.53 / 145.02 ms โ”‚     no change โ”‚
   โ”‚ QQuery 61 โ”‚           14.03 / 14.16 ยฑ0.11 / 14.30 ms โ”‚           13.72 / 13.90 ยฑ0.11 / 14.01 ms โ”‚     no change โ”‚
   โ”‚ QQuery 62 โ”‚        877.31 / 884.77 ยฑ7.38 / 896.32 ms โ”‚        880.19 / 888.63 ยฑ9.39 / 901.53 ms โ”‚     no change โ”‚
   โ”‚ QQuery 63 โ”‚        104.22 / 106.06 ยฑ2.20 / 110.14 ms โ”‚        103.70 / 105.41 ยฑ2.06 / 109.08 ms โ”‚     no change โ”‚
   โ”‚ QQuery 64 โ”‚        645.97 / 651.70 ยฑ4.13 / 657.78 ms โ”‚        645.49 / 653.13 ยฑ9.93 / 672.55 ms โ”‚     no change โ”‚
   โ”‚ QQuery 65 โ”‚        251.72 / 256.12 ยฑ2.82 / 259.37 ms โ”‚        253.25 / 255.82 ยฑ1.86 / 257.82 ms โ”‚     no change โ”‚
   โ”‚ QQuery 66 โ”‚       217.80 / 230.85 ยฑ11.64 / 249.60 ms โ”‚        218.09 / 225.35 ยฑ6.76 / 237.96 ms โ”‚     no change โ”‚
   โ”‚ QQuery 67 โ”‚       296.57 / 308.22 ยฑ10.59 / 325.66 ms โ”‚       300.94 / 314.83 ยฑ14.43 / 340.69 ms โ”‚     no change โ”‚
   โ”‚ QQuery 68 โ”‚              8.75 / 8.97 ยฑ0.20 / 9.34 ms โ”‚              8.78 / 8.99 ยฑ0.13 / 9.14 ms โ”‚     no change โ”‚
   โ”‚ QQuery 69 โ”‚         99.69 / 100.48 ยฑ0.56 / 101.14 ms โ”‚         98.21 / 101.36 ยฑ3.33 / 106.38 ms โ”‚     no change โ”‚
   โ”‚ QQuery 70 โ”‚       317.61 / 329.88 ยฑ11.19 / 349.17 ms โ”‚        316.47 / 326.92 ยฑ9.20 / 340.61 ms โ”‚     no change โ”‚
   โ”‚ QQuery 71 โ”‚        134.79 / 136.84 ยฑ2.15 / 140.75 ms โ”‚        134.71 / 137.70 ยฑ2.99 / 142.32 ms โ”‚     no change โ”‚
   โ”‚ QQuery 72 โ”‚        553.84 / 562.55 ยฑ8.80 / 577.93 ms โ”‚        557.71 / 565.37 ยฑ6.92 / 577.39 ms โ”‚     no change โ”‚
   โ”‚ QQuery 73 โ”‚              7.00 / 7.13 ยฑ0.15 / 7.43 ms โ”‚              6.84 / 7.04 ยฑ0.15 / 7.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 74 โ”‚        602.47 / 612.26 ยฑ6.41 / 620.60 ms โ”‚        615.13 / 618.04 ยฑ1.70 / 619.86 ms โ”‚     no change โ”‚
   โ”‚ QQuery 75 โ”‚        268.97 / 271.69 ยฑ3.64 / 278.90 ms โ”‚        268.38 / 270.66 ยฑ2.39 / 275.06 ms โ”‚     no change โ”‚
   โ”‚ QQuery 76 โ”‚        131.30 / 132.23 ยฑ0.83 / 133.70 ms โ”‚        130.31 / 133.07 ยฑ2.33 / 135.89 ms โ”‚     no change โ”‚
   โ”‚ QQuery 77 โ”‚        189.59 / 190.71 ยฑ1.57 / 193.78 ms โ”‚        188.19 / 189.21 ยฑ1.07 / 191.11 ms โ”‚     no change โ”‚
   โ”‚ QQuery 78 โ”‚        317.08 / 319.14 ยฑ1.98 / 321.52 ms โ”‚        313.56 / 314.54 ยฑ0.81 / 315.40 ms โ”‚     no change โ”‚
   โ”‚ QQuery 79 โ”‚        240.05 / 247.11 ยฑ7.48 / 261.39 ms โ”‚        234.92 / 239.64 ยฑ2.63 / 242.66 ms โ”‚     no change โ”‚
   โ”‚ QQuery 80 โ”‚        303.79 / 307.35 ยฑ3.18 / 313.15 ms โ”‚        302.11 / 303.07 ยฑ0.70 / 304.03 ms โ”‚     no change โ”‚
   โ”‚ QQuery 81 โ”‚           26.47 / 26.88 ยฑ0.42 / 27.59 ms โ”‚           26.27 / 26.69 ยฑ0.32 / 27.18 ms โ”‚     no change โ”‚
   โ”‚ QQuery 82 โ”‚           39.98 / 40.18 ยฑ0.12 / 40.32 ms โ”‚           39.81 / 40.16 ยฑ0.28 / 40.54 ms โ”‚     no change โ”‚
   โ”‚ QQuery 83 โ”‚           36.24 / 36.42 ยฑ0.28 / 36.98 ms โ”‚           35.73 / 36.19 ยฑ0.35 / 36.67 ms โ”‚     no change โ”‚
   โ”‚ QQuery 84 โ”‚           46.76 / 47.51 ยฑ0.97 / 49.43 ms โ”‚           46.38 / 46.71 ยฑ0.39 / 47.46 ms โ”‚     no change โ”‚
   โ”‚ QQuery 85 โ”‚        143.62 / 144.52 ยฑ0.86 / 146.07 ms โ”‚        140.03 / 140.97 ยฑ0.78 / 142.32 ms โ”‚     no change โ”‚
   โ”‚ QQuery 86 โ”‚           38.09 / 38.30 ยฑ0.22 / 38.73 ms โ”‚           37.80 / 39.68 ยฑ2.47 / 44.56 ms โ”‚     no change โ”‚
   โ”‚ QQuery 87 โ”‚              3.65 / 3.78 ยฑ0.14 / 4.05 ms โ”‚              3.64 / 3.75 ยฑ0.21 / 4.18 ms โ”‚     no change โ”‚
   โ”‚ QQuery 88 โ”‚        101.20 / 103.13 ยฑ3.23 / 109.57 ms โ”‚        100.03 / 100.91 ยฑ0.93 / 102.71 ms โ”‚     no change โ”‚
   โ”‚ QQuery 89 โ”‚        119.24 / 119.66 ยฑ0.53 / 120.57 ms โ”‚        117.92 / 121.00 ยฑ4.45 / 129.77 ms โ”‚     no change โ”‚
   โ”‚ QQuery 90 โ”‚           23.01 / 24.60 ยฑ2.41 / 29.41 ms โ”‚           23.30 / 23.71 ยฑ0.34 / 24.11 ms โ”‚     no change โ”‚
   โ”‚ QQuery 91 โ”‚           59.78 / 60.53 ยฑ0.69 / 61.79 ms โ”‚           58.05 / 59.46 ยฑ0.86 / 60.75 ms โ”‚     no change โ”‚
   โ”‚ QQuery 92 โ”‚           57.60 / 58.02 ยฑ0.34 / 58.49 ms โ”‚           57.47 / 58.56 ยฑ1.48 / 61.47 ms โ”‚     no change โ”‚
   โ”‚ QQuery 93 โ”‚        165.50 / 167.21 ยฑ1.22 / 169.01 ms โ”‚        164.59 / 166.85 ยฑ2.24 / 170.57 ms โ”‚     no change โ”‚
   โ”‚ QQuery 94 โ”‚           62.46 / 62.70 ยฑ0.22 / 63.07 ms โ”‚           61.27 / 62.15 ยฑ1.29 / 64.70 ms โ”‚     no change โ”‚
   โ”‚ QQuery 95 โ”‚        114.28 / 115.01 ยฑ0.47 / 115.58 ms โ”‚        112.46 / 113.76 ยฑ0.98 / 115.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 96 โ”‚           69.01 / 69.67 ยฑ0.62 / 70.52 ms โ”‚           67.96 / 70.04 ยฑ1.17 / 71.55 ms โ”‚     no change โ”‚
   โ”‚ QQuery 97 โ”‚        113.42 / 116.55 ยฑ2.46 / 120.69 ms โ”‚        113.10 / 114.98 ยฑ1.03 / 116.23 ms โ”‚     no change โ”‚
   โ”‚ QQuery 98 โ”‚        151.98 / 153.95 ยฑ1.23 / 155.79 ms โ”‚        151.84 / 152.95 ยฑ1.01 / 154.60 ms โ”‚     no change โ”‚
   โ”‚ QQuery 99 โ”‚ 10760.60 / 10829.54 ยฑ69.46 / 10961.65 ms โ”‚ 10806.84 / 10993.60 ยฑ99.37 / 11074.15 ms โ”‚     no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 30548.78ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 30750.67ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚   308.57ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚   310.61ms โ”‚
   โ”‚ Queries Faster                                    โ”‚          2 โ”‚
   โ”‚ Queries Slower                                    โ”‚          1 โ”‚
   โ”‚ Queries with No Change                            โ”‚         96 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **tpcds โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 155.0s |
   | Peak memory | 6.3 GiB |
   | Avg memory | 5.7 GiB |
   | CPU user | 243.2s |
   | CPU sys | 7.9s |
   | Peak spill | 0 B |
   
   **tpcds โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 155.0s |
   | Peak memory | 6.3 GiB |
   | Avg memory | 5.6 GiB |
   | CPU user | 243.4s |
   | CPU sys | 8.9s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364585293

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364579262)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4364579262-1984-fjh5v 6.12.68+ #1 SMP Wed Apr  1 02:23:28 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (877ea557b9b6019f3544956a5d2ddd2c90921730) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..877ea557b9b6019f3544956a5d2ddd2c90921730) using: tpch
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4365837126

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4365794880)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark clickbench_extended.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                      HEAD โ”ƒ        speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 0  โ”‚        780.64 / 810.90 ยฑ18.54 / 837.37 ms โ”‚        881.03 / 931.96 ยฑ33.63 / 974.31 ms โ”‚  1.15x slower โ”‚
   โ”‚ QQuery 1  โ”‚         195.72 / 197.84 ยฑ3.12 / 203.96 ms โ”‚         196.42 / 198.38 ยฑ3.17 / 204.69 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚         480.86 / 484.54 ยฑ2.29 / 488.02 ms โ”‚         476.37 / 479.25 ยฑ2.02 / 481.88 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚         312.84 / 313.95 ยฑ0.73 / 315.09 ms โ”‚         308.98 / 311.48 ยฑ2.73 / 316.67 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚        648.68 / 666.54 ยฑ16.06 / 694.44 ms โ”‚        655.44 / 679.70 ยฑ12.73 / 688.97 ms โ”‚     no change โ”‚
   โ”‚ QQuery 5  โ”‚ 10446.29 / 10712.49 ยฑ212.31 / 10989.87 ms โ”‚ 10348.00 / 10567.24 ยฑ295.63 / 11143.53 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚     1016.61 / 1037.59 ยฑ18.86 / 1066.80 ms โ”‚     1004.58 / 1023.36 ยฑ11.33 / 1039.55 ms โ”‚     no change โ”‚
   โ”‚ QQuery 7  โ”‚        747.68 / 790.13 ยฑ31.97 / 846.11 ms โ”‚         812.63 / 818.51 ยฑ4.86 / 824.35 ms โ”‚     no change โ”‚
   โ”‚ QQuery 8  โ”‚         393.09 / 405.01 ยฑ9.96 / 422.95 ms โ”‚        379.51 / 398.88 ยฑ10.79 / 412.76 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚     2801.41 / 2900.39 ยฑ68.09 / 2995.66 ms โ”‚     2898.76 / 2972.24 ยฑ58.85 / 3075.32 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚        643.63 / 653.01 ยฑ12.49 / 677.69 ms โ”‚         628.57 / 639.30 ยฑ8.57 / 651.40 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚     2119.83 / 2211.61 ยฑ57.38 / 2278.65 ms โ”‚     2196.57 / 2265.52 ยฑ36.17 / 2295.29 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚        192.80 / 212.40 ยฑ22.17 / 253.77 ms โ”‚        186.27 / 196.67 ยฑ12.06 / 220.24 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 13 โ”‚        566.46 / 580.52 ยฑ15.39 / 607.05 ms โ”‚         553.59 / 563.74 ยฑ8.88 / 575.24 ms โ”‚     no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 21976.93ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 22046.21ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚  1569.78ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚  1574.73ms โ”‚
   โ”‚ Queries Faster                                    โ”‚          1 โ”‚
   โ”‚ Queries Slower                                    โ”‚          1 โ”‚
   โ”‚ Queries with No Change                            โ”‚         12 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **clickbench_extended โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 115.0s |
   | Peak memory | 30.1 GiB |
   | Avg memory | 23.7 GiB |
   | CPU user | 1064.6s |
   | CPU sys | 43.6s |
   | Peak spill | 0 B |
   
   **clickbench_extended โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 115.0s |
   | Peak memory | 31.3 GiB |
   | Avg memory | 23.8 GiB |
   | CPU user | 1056.5s |
   | CPU sys | 43.3s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4365799786

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4365794880)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4365794880-1985-dr9sn 6.12.68+ #1 SMP Wed Apr  1 02:23:28 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (902da95a843fc11f42ad3f8182de66dfdd2488ce) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..902da95a843fc11f42ad3f8182de66dfdd2488ce) using: clickbench_extended
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363673777

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4363668679)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4363668679-1974-gcmkk 6.12.55+ #1 SMP Sun Feb  1 08:59:41 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (58beeff7c2416a3e4da2644b320ff57f4c3e4016) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..58beeff7c2416a3e4da2644b320ff57f4c3e4016) using: clickbench_partitioned
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4361073956

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4360963663)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark clickbench_partitioned.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                  HEAD โ”ƒ    speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 0  โ”‚          1.18 / 4.58 ยฑ6.69 / 17.96 ms โ”‚          1.19 / 4.62 ยฑ6.73 / 18.07 ms โ”‚     no change โ”‚
   โ”‚ QQuery 1  โ”‚        12.28 / 12.67 ยฑ0.23 / 12.97 ms โ”‚        12.51 / 12.61 ยฑ0.07 / 12.72 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚        36.79 / 37.05 ยฑ0.16 / 37.26 ms โ”‚        36.53 / 37.20 ยฑ0.44 / 37.92 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚        31.65 / 32.53 ยฑ1.49 / 35.51 ms โ”‚        32.04 / 32.97 ยฑ0.53 / 33.57 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚     230.83 / 236.51 ยฑ6.05 / 248.20 ms โ”‚     214.37 / 215.96 ยฑ1.58 / 218.54 ms โ”‚ +1.10x faster โ”‚
   โ”‚ QQuery 5  โ”‚     281.64 / 282.89 ยฑ1.10 / 284.36 ms โ”‚     279.04 / 281.26 ยฑ1.31 / 283.04 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚           6.52 / 7.41 ยฑ0.73 / 8.64 ms โ”‚           6.82 / 7.45 ยฑ0.73 / 8.88 ms โ”‚     no change โ”‚
   โ”‚ QQuery 7  โ”‚        13.36 / 13.57 ยฑ0.13 / 13.73 ms โ”‚        13.63 / 13.71 ยฑ0.07 / 13.82 ms โ”‚     no change โ”‚
   โ”‚ QQuery 8  โ”‚     318.97 / 321.36 ยฑ1.46 / 323.55 ms โ”‚     313.37 / 317.26 ยฑ2.52 / 320.90 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚     450.81 / 458.18 ยฑ4.63 / 464.33 ms โ”‚     443.68 / 450.71 ยฑ5.57 / 459.18 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚        73.89 / 74.59 ยฑ0.63 / 75.64 ms โ”‚        73.18 / 73.91 ยฑ0.63 / 74.94 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚        84.39 / 85.54 ยฑ0.83 / 86.55 ms โ”‚        84.65 / 86.23 ยฑ1.03 / 87.49 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚     273.58 / 278.49 ยฑ3.98 / 285.24 ms โ”‚     270.37 / 274.69 ยฑ4.45 / 280.94 ms โ”‚     no change โ”‚
   โ”‚ QQuery 13 โ”‚     385.80 / 390.52 ยฑ4.98 / 399.96 ms โ”‚     377.50 / 383.73 ยฑ4.73 / 389.87 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚     286.47 / 287.76 ยฑ0.81 / 289.03 ms โ”‚     280.90 / 286.96 ยฑ5.70 / 295.80 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚     275.97 / 280.29 ยฑ2.77 / 284.33 ms โ”‚     260.26 / 263.90 ยฑ2.20 / 265.85 ms โ”‚ +1.06x faster โ”‚
   โ”‚ QQuery 16 โ”‚     612.13 / 618.96 ยฑ5.27 / 628.09 ms โ”‚    614.42 / 622.50 ยฑ10.54 / 641.61 ms โ”‚     no change โ”‚
   โ”‚ QQuery 17 โ”‚     611.92 / 618.02 ยฑ5.35 / 627.10 ms โ”‚     607.58 / 615.65 ยฑ4.98 / 622.96 ms โ”‚     no change โ”‚
   โ”‚ QQuery 18 โ”‚  1235.82 / 1246.71 ยฑ9.88 / 1265.11 ms โ”‚ 1224.87 / 1240.38 ยฑ12.24 / 1255.12 ms โ”‚     no change โ”‚
   โ”‚ QQuery 19 โ”‚       29.14 / 39.85 ยฑ10.47 / 57.16 ms โ”‚       28.66 / 38.02 ยฑ10.40 / 53.24 ms โ”‚     no change โ”‚
   โ”‚ QQuery 20 โ”‚     518.86 / 527.17 ยฑ7.71 / 537.16 ms โ”‚     523.03 / 529.19 ยฑ5.51 / 537.33 ms โ”‚     no change โ”‚
   โ”‚ QQuery 21 โ”‚     600.66 / 603.37 ยฑ2.34 / 606.32 ms โ”‚     597.33 / 600.73 ยฑ3.25 / 606.75 ms โ”‚     no change โ”‚
   โ”‚ QQuery 22 โ”‚  1065.42 / 1072.81 ยฑ5.71 / 1080.54 ms โ”‚ 1066.09 / 1088.44 ยฑ16.48 / 1113.93 ms โ”‚     no change โ”‚
   โ”‚ QQuery 23 โ”‚ 3360.41 / 3390.06 ยฑ15.86 / 3407.55 ms โ”‚ 3359.93 / 3381.70 ยฑ14.96 / 3401.24 ms โ”‚     no change โ”‚
   โ”‚ QQuery 24 โ”‚        42.18 / 44.34 ยฑ3.81 / 51.96 ms โ”‚        41.88 / 45.71 ยฑ4.91 / 55.17 ms โ”‚     no change โ”‚
   โ”‚ QQuery 25 โ”‚     113.47 / 116.59 ยฑ4.04 / 124.56 ms โ”‚     113.89 / 114.56 ยฑ0.41 / 114.93 ms โ”‚     no change โ”‚
   โ”‚ QQuery 26 โ”‚        43.08 / 43.71 ยฑ0.59 / 44.46 ms โ”‚        42.92 / 43.63 ยฑ0.62 / 44.76 ms โ”‚     no change โ”‚
   โ”‚ QQuery 27 โ”‚     668.70 / 675.38 ยฑ4.10 / 680.57 ms โ”‚     670.18 / 682.75 ยฑ9.76 / 697.34 ms โ”‚     no change โ”‚
   โ”‚ QQuery 28 โ”‚  3005.24 / 3013.95 ยฑ6.55 / 3023.19 ms โ”‚ 3011.19 / 3022.17 ยฑ13.72 / 3048.41 ms โ”‚     no change โ”‚
   โ”‚ QQuery 29 โ”‚        42.44 / 45.11 ยฑ4.18 / 53.41 ms โ”‚        42.34 / 46.77 ยฑ7.22 / 61.11 ms โ”‚     no change โ”‚
   โ”‚ QQuery 30 โ”‚     306.47 / 313.41 ยฑ6.20 / 325.08 ms โ”‚     307.38 / 318.80 ยฑ9.22 / 333.64 ms โ”‚     no change โ”‚
   โ”‚ QQuery 31 โ”‚     298.52 / 305.86 ยฑ6.19 / 315.89 ms โ”‚     301.85 / 305.33 ยฑ3.19 / 309.99 ms โ”‚     no change โ”‚
   โ”‚ QQuery 32 โ”‚   974.28 / 988.73 ยฑ12.42 / 1008.28 ms โ”‚    963.69 / 978.29 ยฑ12.27 / 992.08 ms โ”‚     no change โ”‚
   โ”‚ QQuery 33 โ”‚ 1420.14 / 1444.11 ยฑ13.82 / 1460.16 ms โ”‚  1429.94 / 1435.09 ยฑ6.62 / 1448.19 ms โ”‚     no change โ”‚
   โ”‚ QQuery 34 โ”‚ 1427.78 / 1450.88 ยฑ12.92 / 1465.79 ms โ”‚  1442.66 / 1452.02 ยฑ9.34 / 1469.85 ms โ”‚     no change โ”‚
   โ”‚ QQuery 35 โ”‚    278.26 / 291.57 ยฑ12.11 / 312.83 ms โ”‚    237.48 / 250.81 ยฑ10.62 / 269.68 ms โ”‚ +1.16x faster โ”‚
   โ”‚ QQuery 36 โ”‚        61.58 / 68.17 ยฑ5.38 / 75.42 ms โ”‚        62.23 / 63.35 ยฑ0.95 / 64.79 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 37 โ”‚        36.02 / 39.54 ยฑ2.27 / 42.87 ms โ”‚        35.12 / 39.01 ยฑ4.92 / 48.28 ms โ”‚     no change โ”‚
   โ”‚ QQuery 38 โ”‚        40.98 / 44.98 ยฑ3.84 / 52.03 ms โ”‚        43.62 / 46.67 ยฑ3.23 / 52.05 ms โ”‚     no change โ”‚
   โ”‚ QQuery 39 โ”‚     123.78 / 132.74 ยฑ5.43 / 140.23 ms โ”‚     128.25 / 136.81 ยฑ6.35 / 147.04 ms โ”‚     no change โ”‚
   โ”‚ QQuery 40 โ”‚        14.70 / 19.44 ยฑ4.08 / 26.14 ms โ”‚        14.21 / 14.70 ยฑ0.73 / 16.13 ms โ”‚ +1.32x faster โ”‚
   โ”‚ QQuery 41 โ”‚        14.02 / 14.36 ยฑ0.42 / 15.17 ms โ”‚        13.93 / 14.79 ยฑ1.14 / 16.99 ms โ”‚     no change โ”‚
   โ”‚ QQuery 42 โ”‚        13.42 / 13.48 ยฑ0.04 / 13.54 ms โ”‚        13.19 / 15.87 ยฑ3.81 / 23.19 ms โ”‚  1.18x slower โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 19987.21ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 19886.90ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚   464.82ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚   462.49ms โ”‚
   โ”‚ Queries Faster                                    โ”‚          5 โ”‚
   โ”‚ Queries Slower                                    โ”‚          1 โ”‚
   โ”‚ Queries with No Change                            โ”‚         37 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **clickbench_partitioned โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 105.0s |
   | Peak memory | 29.4 GiB |
   | Avg memory | 22.9 GiB |
   | CPU user | 1055.4s |
   | CPU sys | 66.0s |
   | Peak spill | 0 B |
   
   **clickbench_partitioned โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 100.0s |
   | Peak memory | 30.6 GiB |
   | Avg memory | 23.4 GiB |
   | CPU user | 1049.7s |
   | CPU sys | 65.1s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364579262

   run benchmarks


-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363628529

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4363624010)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4363624010-1971-sx4lk 6.12.55+ #1 SMP Sun Feb  1 08:59:41 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (58beeff7c2416a3e4da2644b320ff57f4c3e4016) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..58beeff7c2416a3e4da2644b320ff57f4c3e4016) using: tpcds
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364124826

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364085635)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark clickbench_partitioned.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                  HEAD โ”ƒ    speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 0  โ”‚          1.19 / 4.54 ยฑ6.64 / 17.83 ms โ”‚          1.18 / 4.55 ยฑ6.68 / 17.92 ms โ”‚     no change โ”‚
   โ”‚ QQuery 1  โ”‚        12.36 / 12.61 ยฑ0.13 / 12.73 ms โ”‚        12.38 / 12.50 ยฑ0.08 / 12.62 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚        36.39 / 36.67 ยฑ0.30 / 37.12 ms โ”‚        36.61 / 36.88 ยฑ0.23 / 37.15 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚        31.43 / 32.04 ยฑ0.55 / 33.08 ms โ”‚        31.63 / 31.99 ยฑ0.39 / 32.65 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚     232.41 / 235.55 ยฑ2.52 / 238.49 ms โ”‚     227.00 / 229.43 ยฑ1.86 / 232.19 ms โ”‚     no change โ”‚
   โ”‚ QQuery 5  โ”‚     279.71 / 282.76 ยฑ2.15 / 285.09 ms โ”‚     276.36 / 279.27 ยฑ1.78 / 281.04 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚           6.05 / 7.49 ยฑ0.83 / 8.57 ms โ”‚           6.43 / 7.21 ยฑ0.68 / 8.34 ms โ”‚     no change โ”‚
   โ”‚ QQuery 7  โ”‚        13.20 / 13.43 ยฑ0.15 / 13.59 ms โ”‚        13.48 / 13.61 ยฑ0.12 / 13.77 ms โ”‚     no change โ”‚
   โ”‚ QQuery 8  โ”‚     316.06 / 319.12 ยฑ2.60 / 323.43 ms โ”‚     314.18 / 316.22 ยฑ1.24 / 317.60 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚     438.66 / 450.07 ยฑ7.45 / 458.24 ms โ”‚     440.88 / 450.43 ยฑ4.85 / 454.36 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚        72.78 / 74.18 ยฑ1.31 / 76.54 ms โ”‚        73.16 / 73.34 ยฑ0.15 / 73.57 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚        84.50 / 86.35 ยฑ2.97 / 92.24 ms โ”‚        82.61 / 83.64 ยฑ0.97 / 85.06 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚     269.07 / 273.09 ยฑ2.63 / 276.72 ms โ”‚     267.57 / 271.75 ยฑ3.33 / 275.64 ms โ”‚     no change โ”‚
   โ”‚ QQuery 13 โ”‚     382.08 / 389.69 ยฑ6.10 / 399.21 ms โ”‚     380.39 / 388.24 ยฑ6.27 / 395.83 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚     281.76 / 284.14 ยฑ2.58 / 288.96 ms โ”‚     279.61 / 285.82 ยฑ6.07 / 295.13 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚    274.14 / 282.46 ยฑ11.29 / 304.85 ms โ”‚     269.69 / 279.84 ยฑ7.79 / 293.09 ms โ”‚     no change โ”‚
   โ”‚ QQuery 16 โ”‚     609.57 / 614.53 ยฑ4.66 / 620.77 ms โ”‚     610.28 / 614.35 ยฑ2.35 / 617.50 ms โ”‚     no change โ”‚
   โ”‚ QQuery 17 โ”‚     608.50 / 615.07 ยฑ4.32 / 619.85 ms โ”‚     609.90 / 614.24 ยฑ3.10 / 617.83 ms โ”‚     no change โ”‚
   โ”‚ QQuery 18 โ”‚  1219.24 / 1231.25 ยฑ7.33 / 1240.30 ms โ”‚ 1221.02 / 1237.05 ยฑ10.41 / 1251.83 ms โ”‚     no change โ”‚
   โ”‚ QQuery 19 โ”‚        28.56 / 29.00 ยฑ0.47 / 29.85 ms โ”‚        28.61 / 32.05 ยฑ4.97 / 41.54 ms โ”‚  1.11x slower โ”‚
   โ”‚ QQuery 20 โ”‚     522.87 / 527.19 ยฑ5.20 / 536.66 ms โ”‚    519.55 / 528.55 ยฑ13.65 / 555.57 ms โ”‚     no change โ”‚
   โ”‚ QQuery 21 โ”‚     593.94 / 602.78 ยฑ5.04 / 609.10 ms โ”‚     597.92 / 603.02 ยฑ4.99 / 609.88 ms โ”‚     no change โ”‚
   โ”‚ QQuery 22 โ”‚ 1058.00 / 1071.63 ยฑ10.99 / 1085.30 ms โ”‚ 1062.73 / 1076.20 ยฑ12.16 / 1097.11 ms โ”‚     no change โ”‚
   โ”‚ QQuery 23 โ”‚ 3351.92 / 3373.71 ยฑ15.73 / 3392.14 ms โ”‚  3328.17 / 3342.26 ยฑ9.31 / 3355.46 ms โ”‚     no change โ”‚
   โ”‚ QQuery 24 โ”‚        42.04 / 42.94 ยฑ0.93 / 44.61 ms โ”‚        41.95 / 44.50 ยฑ4.92 / 54.33 ms โ”‚     no change โ”‚
   โ”‚ QQuery 25 โ”‚     114.26 / 116.39 ยฑ2.89 / 122.09 ms โ”‚     112.94 / 119.82 ยฑ7.57 / 133.95 ms โ”‚     no change โ”‚
   โ”‚ QQuery 26 โ”‚        43.16 / 43.71 ยฑ0.47 / 44.27 ms โ”‚        42.50 / 43.33 ยฑ1.04 / 45.38 ms โ”‚     no change โ”‚
   โ”‚ QQuery 27 โ”‚     666.54 / 677.07 ยฑ6.67 / 683.88 ms โ”‚     669.74 / 677.46 ยฑ6.39 / 686.41 ms โ”‚     no change โ”‚
   โ”‚ QQuery 28 โ”‚  2985.63 / 2995.70 ยฑ8.30 / 3008.74 ms โ”‚ 3010.93 / 3032.88 ยฑ18.51 / 3064.09 ms โ”‚     no change โ”‚
   โ”‚ QQuery 29 โ”‚        42.67 / 46.29 ยฑ5.75 / 57.58 ms โ”‚       42.00 / 48.50 ยฑ12.56 / 73.62 ms โ”‚     no change โ”‚
   โ”‚ QQuery 30 โ”‚     303.32 / 307.24 ยฑ2.97 / 311.79 ms โ”‚     303.69 / 309.49 ยฑ5.95 / 318.55 ms โ”‚     no change โ”‚
   โ”‚ QQuery 31 โ”‚     299.13 / 304.25 ยฑ4.55 / 311.61 ms โ”‚     301.44 / 303.38 ยฑ2.42 / 308.08 ms โ”‚     no change โ”‚
   โ”‚ QQuery 32 โ”‚    968.57 / 982.21 ยฑ10.40 / 996.80 ms โ”‚   959.32 / 980.79 ยฑ21.08 / 1017.01 ms โ”‚     no change โ”‚
   โ”‚ QQuery 33 โ”‚  1419.93 / 1425.16 ยฑ4.00 / 1431.79 ms โ”‚ 1406.08 / 1423.09 ยฑ12.22 / 1437.25 ms โ”‚     no change โ”‚
   โ”‚ QQuery 34 โ”‚  1427.46 / 1442.94 ยฑ9.65 / 1453.18 ms โ”‚ 1428.77 / 1449.49 ยฑ27.15 / 1502.62 ms โ”‚     no change โ”‚
   โ”‚ QQuery 35 โ”‚     278.77 / 283.02 ยฑ2.86 / 286.44 ms โ”‚    263.91 / 282.01 ยฑ16.26 / 300.79 ms โ”‚     no change โ”‚
   โ”‚ QQuery 36 โ”‚        60.91 / 65.43 ยฑ3.66 / 68.85 ms โ”‚        60.88 / 65.17 ยฑ3.70 / 70.94 ms โ”‚     no change โ”‚
   โ”‚ QQuery 37 โ”‚        35.01 / 36.25 ยฑ0.77 / 37.00 ms โ”‚        34.47 / 35.97 ยฑ2.09 / 40.06 ms โ”‚     no change โ”‚
   โ”‚ QQuery 38 โ”‚        39.65 / 41.64 ยฑ1.22 / 43.28 ms โ”‚        40.93 / 42.96 ยฑ1.88 / 46.23 ms โ”‚     no change โ”‚
   โ”‚ QQuery 39 โ”‚     128.32 / 131.62 ยฑ2.65 / 135.03 ms โ”‚     118.87 / 130.09 ยฑ7.31 / 141.83 ms โ”‚     no change โ”‚
   โ”‚ QQuery 40 โ”‚        13.87 / 14.05 ยฑ0.13 / 14.21 ms โ”‚        13.87 / 14.38 ยฑ0.79 / 15.94 ms โ”‚     no change โ”‚
   โ”‚ QQuery 41 โ”‚        13.29 / 14.67 ยฑ1.93 / 18.50 ms โ”‚        13.44 / 13.63 ยฑ0.13 / 13.85 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 42 โ”‚        13.01 / 13.99 ยฑ1.59 / 17.16 ms โ”‚        13.57 / 15.88 ยฑ4.21 / 24.30 ms โ”‚  1.13x slower โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 19833.90ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 19845.27ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚   461.25ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚   461.52ms โ”‚
   โ”‚ Queries Faster                                    โ”‚          1 โ”‚
   โ”‚ Queries Slower                                    โ”‚          2 โ”‚
   โ”‚ Queries with No Change                            โ”‚         40 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **clickbench_partitioned โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 100.0s |
   | Peak memory | 30.7 GiB |
   | Avg memory | 23.5 GiB |
   | CPU user | 1051.8s |
   | CPU sys | 62.4s |
   | Peak spill | 0 B |
   
   **clickbench_partitioned โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 100.0s |
   | Peak memory | 30.1 GiB |
   | Avg memory | 23.4 GiB |
   | CPU user | 1047.1s |
   | CPU sys | 63.5s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364091674

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364085869)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4364085869-1978-bztjf 6.12.68+ #1 SMP Wed Apr  1 02:23:28 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (08629a74dc3ee2842f44e43fe8ee96f8c3af7dea) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..08629a74dc3ee2842f44e43fe8ee96f8c3af7dea) using: clickbench_extended
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363651219

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4363624010)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark tpch_sf1.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                           HEAD โ”ƒ speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 1  โ”‚ 41.01 / 42.21 ยฑ0.83 / 43.41 ms โ”‚     40.85 / 41.64 ยฑ1.09 / 43.67 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚ 21.38 / 21.55 ยฑ0.29 / 22.12 ms โ”‚     21.52 / 22.44 ยฑ1.03 / 23.70 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚ 36.90 / 39.89 ยฑ1.92 / 42.95 ms โ”‚     40.24 / 41.14 ยฑ1.04 / 42.78 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚ 18.77 / 19.31 ยฑ0.70 / 20.65 ms โ”‚     18.62 / 18.83 ยฑ0.18 / 19.07 ms โ”‚     no change โ”‚
   โ”‚ QQuery 5  โ”‚ 45.85 / 46.89 ยฑ0.97 / 48.42 ms โ”‚     44.82 / 46.80 ยฑ1.39 / 48.14 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚ 17.97 / 18.74 ยฑ1.02 / 20.74 ms โ”‚     17.87 / 18.69 ยฑ1.02 / 20.63 ms โ”‚     no change โ”‚
   โ”‚ QQuery 7  โ”‚ 53.87 / 55.92 ยฑ1.24 / 57.56 ms โ”‚     51.85 / 53.62 ยฑ1.35 / 55.50 ms โ”‚     no change โ”‚
   โ”‚ QQuery 8  โ”‚ 48.57 / 49.08 ยฑ0.43 / 49.71 ms โ”‚     47.94 / 48.43 ยฑ0.38 / 48.90 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚ 53.28 / 54.07 ยฑ0.72 / 55.42 ms โ”‚     54.03 / 55.19 ยฑ1.32 / 57.68 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚ 68.32 / 69.25 ยฑ0.96 / 70.43 ms โ”‚     68.37 / 69.01 ยฑ1.16 / 71.33 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚ 14.51 / 14.76 ยฑ0.18 / 15.08 ms โ”‚     14.92 / 15.18 ยฑ0.21 / 15.52 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚ 26.91 / 27.08 ยฑ0.13 / 27.26 ms โ”‚     27.59 / 28.09 ยฑ0.55 / 29.13 ms โ”‚     no change โ”‚
   โ”‚ QQuery 13 โ”‚ 36.23 / 37.08 ยฑ0.60 / 37.79 ms โ”‚     36.50 / 37.39 ยฑ0.56 / 38.23 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚ 28.20 / 28.68 ยฑ0.59 / 29.74 ms โ”‚     27.62 / 28.65 ยฑ1.25 / 31.09 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚ 35.51 / 36.80 ยฑ0.97 / 37.99 ms โ”‚     34.19 / 34.74 ยฑ0.65 / 35.98 ms โ”‚ +1.06x faster โ”‚
   โ”‚ QQuery 16 โ”‚ 16.27 / 16.50 ยฑ0.15 / 16.67 ms โ”‚     15.77 / 15.85 ยฑ0.12 / 16.08 ms โ”‚     no change โ”‚
   โ”‚ QQuery 17 โ”‚ 82.48 / 84.17 ยฑ1.32 / 85.85 ms โ”‚     71.88 / 72.85 ยฑ0.64 / 73.81 ms โ”‚ +1.16x faster โ”‚
   โ”‚ QQuery 18 โ”‚ 74.15 / 76.28 ยฑ2.27 / 79.74 ms โ”‚     70.68 / 71.16 ยฑ0.24 / 71.36 ms โ”‚ +1.07x faster โ”‚
   โ”‚ QQuery 19 โ”‚ 39.33 / 39.94 ยฑ0.47 / 40.59 ms โ”‚     38.22 / 38.53 ยฑ0.24 / 38.94 ms โ”‚     no change โ”‚
   โ”‚ QQuery 20 โ”‚ 41.12 / 41.45 ยฑ0.26 / 41.86 ms โ”‚     39.56 / 39.65 ยฑ0.10 / 39.83 ms โ”‚     no change โ”‚
   โ”‚ QQuery 21 โ”‚ 62.17 / 63.81 ยฑ1.20 / 65.58 ms โ”‚     60.60 / 63.29 ยฑ3.33 / 69.79 ms โ”‚     no change โ”‚
   โ”‚ QQuery 22 โ”‚ 24.99 / 25.91 ยฑ0.98 / 27.47 ms โ”‚     24.58 / 24.90 ยฑ0.36 / 25.58 ms โ”‚     no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ          โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 909.39ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 886.08ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚  41.34ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚  40.28ms โ”‚
   โ”‚ Queries Faster                                    โ”‚        3 โ”‚
   โ”‚ Queries Slower                                    โ”‚        0 โ”‚
   โ”‚ Queries with No Change                            โ”‚       19 โ”‚
   โ”‚ Queries with Failure                              โ”‚        0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **tpch โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 5.0s |
   | Peak memory | 5.8 GiB |
   | Avg memory | 5.1 GiB |
   | CPU user | 33.8s |
   | CPU sys | 2.5s |
   | Peak spill | 0 B |
   
   **tpch โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 5.0s |
   | Peak memory | 5.4 GiB |
   | Avg memory | 5.0 GiB |
   | CPU user | 32.6s |
   | CPU sys | 2.6s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363628456

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4363624206)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4363624206-1973-nk52z 6.12.55+ #1 SMP Sun Feb  1 08:59:41 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (58beeff7c2416a3e4da2644b320ff57f4c3e4016) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..58beeff7c2416a3e4da2644b320ff57f4c3e4016) using: clickbench_extended
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363624206

   run benchmark clickbench_extended


-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4365799431

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4365794880)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4365794880-1986-fsmf7 6.12.68+ #1 SMP Wed Apr  1 02:23:28 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (902da95a843fc11f42ad3f8182de66dfdd2488ce) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..902da95a843fc11f42ad3f8182de66dfdd2488ce) using: tpch10
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4361073273

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4360963663)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark tpcds_sf1.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                      HEAD โ”ƒ       speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 1  โ”‚               7.05 / 7.64 ยฑ0.79 / 9.21 ms โ”‚              7.49 / 7.91 ยฑ0.76 / 9.42 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚         144.63 / 145.18 ยฑ0.38 / 145.78 ms โ”‚        145.01 / 145.37 ยฑ0.25 / 145.72 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚         114.97 / 116.59 ยฑ1.16 / 118.50 ms โ”‚        115.36 / 115.84 ยฑ0.72 / 117.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚     1292.86 / 1385.98 ยฑ85.68 / 1488.41 ms โ”‚    1258.09 / 1280.51 ยฑ18.46 / 1303.64 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 5  โ”‚         173.25 / 174.04 ยฑ1.35 / 176.73 ms โ”‚        172.38 / 173.40 ยฑ0.90 / 174.96 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚         136.71 / 139.43 ยฑ2.11 / 142.10 ms โ”‚        128.46 / 132.45 ยฑ2.23 / 135.35 ms โ”‚ +1.05x faster โ”‚
   โ”‚ QQuery 7  โ”‚         319.17 / 321.14 ยฑ1.76 / 323.97 ms โ”‚        317.17 / 318.33 ยฑ1.15 / 320.04 ms โ”‚     no change โ”‚
   โ”‚ QQuery 8  โ”‚         112.30 / 112.59 ยฑ0.40 / 113.39 ms โ”‚        111.76 / 113.06 ยฑ0.89 / 114.47 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚          97.34 / 104.57 ยฑ8.81 / 121.86 ms โ”‚         99.23 / 104.04 ยฑ8.43 / 120.86 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚         101.81 / 103.06 ยฑ1.35 / 105.58 ms โ”‚        102.19 / 103.62 ยฑ1.84 / 107.23 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚       898.54 / 988.14 ยฑ64.87 / 1070.20 ms โ”‚      858.57 / 951.56 ยฑ76.73 / 1043.72 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚            43.46 / 44.01 ยฑ0.49 / 44.94 ms โ”‚           45.79 / 46.22 ยฑ0.30 / 46.70 ms โ”‚  1.05x slower โ”‚
   โ”‚ QQuery 13 โ”‚         390.23 / 392.66 ยฑ2.50 / 397.45 ms โ”‚        391.40 / 402.85 ยฑ8.99 / 411.38 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚      1028.41 / 1031.51 ยฑ2.05 / 1033.92 ms โ”‚     1020.04 / 1022.95 ยฑ2.56 / 1027.23 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚            14.28 / 14.78 ยฑ0.32 / 15.18 ms โ”‚           14.20 / 14.61 ยฑ0.39 / 15.14 ms โ”‚     no change โ”‚
   โ”‚ QQuery 16 โ”‚               7.26 / 7.38 ยฑ0.17 / 7.70 ms โ”‚              7.44 / 7.54 ยฑ0.13 / 7.79 ms โ”‚     no change โ”‚
   โ”‚ QQuery 17 โ”‚         201.07 / 202.37 ยฑ1.21 / 204.50 ms โ”‚        201.70 / 208.75 ยฑ4.60 / 213.00 ms โ”‚     no change โ”‚
   โ”‚ QQuery 18 โ”‚         123.49 / 124.15 ยฑ0.60 / 125.04 ms โ”‚        127.63 / 129.23 ยฑ1.32 / 131.16 ms โ”‚     no change โ”‚
   โ”‚ QQuery 19 โ”‚         153.03 / 155.36 ยฑ2.35 / 159.85 ms โ”‚        163.51 / 164.80 ยฑ0.96 / 166.47 ms โ”‚  1.06x slower โ”‚
   โ”‚ QQuery 20 โ”‚            13.06 / 13.33 ยฑ0.29 / 13.74 ms โ”‚           13.97 / 14.14 ยฑ0.09 / 14.24 ms โ”‚  1.06x slower โ”‚
   โ”‚ QQuery 21 โ”‚            18.94 / 19.16 ยฑ0.12 / 19.29 ms โ”‚           19.99 / 20.30 ยฑ0.32 / 20.79 ms โ”‚  1.06x slower โ”‚
   โ”‚ QQuery 22 โ”‚         473.22 / 476.03 ยฑ1.98 / 478.35 ms โ”‚       474.36 / 496.69 ยฑ22.54 / 525.50 ms โ”‚     no change โ”‚
   โ”‚ QQuery 23 โ”‚     1032.34 / 1113.54 ยฑ40.96 / 1140.56 ms โ”‚    1003.84 / 1023.86 ยฑ20.88 / 1061.09 ms โ”‚ +1.09x faster โ”‚
   โ”‚ QQuery 24 โ”‚         631.88 / 634.91 ยฑ4.09 / 642.83 ms โ”‚        630.95 / 633.20 ยฑ1.62 / 635.96 ms โ”‚     no change โ”‚
   โ”‚ QQuery 25 โ”‚         304.13 / 305.95 ยฑ2.19 / 310.27 ms โ”‚        303.46 / 306.93 ยฑ2.27 / 310.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 26 โ”‚            75.92 / 76.48 ยฑ0.41 / 77.01 ms โ”‚           76.64 / 77.30 ยฑ0.48 / 77.73 ms โ”‚     no change โ”‚
   โ”‚ QQuery 27 โ”‚               6.94 / 7.11 ยฑ0.19 / 7.46 ms โ”‚             7.14 / 7.85 ยฑ1.18 / 10.20 ms โ”‚  1.10x slower โ”‚
   โ”‚ QQuery 28 โ”‚         146.55 / 148.00 ยฑ1.11 / 149.88 ms โ”‚        149.51 / 151.82 ยฑ1.97 / 154.04 ms โ”‚     no change โ”‚
   โ”‚ QQuery 29 โ”‚         242.78 / 246.34 ยฑ3.01 / 251.73 ms โ”‚        249.09 / 260.90 ยฑ6.48 / 268.93 ms โ”‚  1.06x slower โ”‚
   โ”‚ QQuery 30 โ”‚            41.16 / 42.14 ยฑ0.97 / 43.85 ms โ”‚           41.94 / 42.50 ยฑ0.56 / 43.51 ms โ”‚     no change โ”‚
   โ”‚ QQuery 31 โ”‚         162.50 / 164.33 ยฑ1.25 / 166.16 ms โ”‚        166.86 / 174.62 ยฑ6.33 / 185.21 ms โ”‚  1.06x slower โ”‚
   โ”‚ QQuery 32 โ”‚            13.11 / 13.35 ยฑ0.30 / 13.94 ms โ”‚           14.87 / 15.26 ยฑ0.25 / 15.62 ms โ”‚  1.14x slower โ”‚
   โ”‚ QQuery 33 โ”‚         138.10 / 140.50 ยฑ2.21 / 144.10 ms โ”‚        147.22 / 148.67 ยฑ1.12 / 150.36 ms โ”‚  1.06x slower โ”‚
   โ”‚ QQuery 34 โ”‚               6.97 / 7.14 ยฑ0.11 / 7.28 ms โ”‚             7.90 / 8.63 ยฑ1.23 / 11.08 ms โ”‚  1.21x slower โ”‚
   โ”‚ QQuery 35 โ”‚          99.80 / 101.63 ยฑ1.61 / 103.81 ms โ”‚        100.92 / 103.60 ยฑ3.11 / 109.68 ms โ”‚     no change โ”‚
   โ”‚ QQuery 36 โ”‚               6.77 / 6.99 ยฑ0.14 / 7.22 ms โ”‚              6.65 / 6.96 ยฑ0.20 / 7.17 ms โ”‚     no change โ”‚
   โ”‚ QQuery 37 โ”‚               8.33 / 8.39 ยฑ0.04 / 8.45 ms โ”‚              8.27 / 8.39 ยฑ0.08 / 8.48 ms โ”‚     no change โ”‚
   โ”‚ QQuery 38 โ”‚            84.47 / 85.04 ยฑ0.44 / 85.59 ms โ”‚           86.06 / 88.84 ยฑ3.41 / 95.12 ms โ”‚     no change โ”‚
   โ”‚ QQuery 39 โ”‚         116.10 / 119.70 ยฑ4.47 / 128.21 ms โ”‚        117.92 / 120.52 ยฑ3.00 / 126.25 ms โ”‚     no change โ”‚
   โ”‚ QQuery 40 โ”‚            93.63 / 95.22 ยฑ1.30 / 97.10 ms โ”‚          94.30 / 98.07 ยฑ2.89 / 101.95 ms โ”‚     no change โ”‚
   โ”‚ QQuery 41 โ”‚            14.24 / 14.47 ยฑ0.28 / 15.01 ms โ”‚           14.77 / 14.99 ยฑ0.17 / 15.20 ms โ”‚     no change โ”‚
   โ”‚ QQuery 42 โ”‚         107.10 / 107.83 ยฑ1.07 / 109.96 ms โ”‚        107.23 / 108.91 ยฑ1.92 / 112.64 ms โ”‚     no change โ”‚
   โ”‚ QQuery 43 โ”‚               5.83 / 6.29 ยฑ0.70 / 7.66 ms โ”‚              5.75 / 5.89 ยฑ0.21 / 6.30 ms โ”‚ +1.07x faster โ”‚
   โ”‚ QQuery 44 โ”‚            10.78 / 10.91 ยฑ0.10 / 11.05 ms โ”‚           10.88 / 11.02 ยฑ0.13 / 11.23 ms โ”‚     no change โ”‚
   โ”‚ QQuery 45 โ”‚            44.57 / 45.02 ยฑ0.45 / 45.88 ms โ”‚           44.21 / 45.68 ยฑ2.22 / 50.06 ms โ”‚     no change โ”‚
   โ”‚ QQuery 46 โ”‚               8.43 / 8.58 ยฑ0.17 / 8.89 ms โ”‚              8.52 / 8.70 ยฑ0.20 / 9.08 ms โ”‚     no change โ”‚
   โ”‚ QQuery 47 โ”‚        693.43 / 793.97 ยฑ69.01 / 860.59 ms โ”‚        670.99 / 681.03 ยฑ6.33 / 688.95 ms โ”‚ +1.17x faster โ”‚
   โ”‚ QQuery 48 โ”‚         286.50 / 295.28 ยฑ8.81 / 311.53 ms โ”‚        270.54 / 274.18 ยฑ2.94 / 278.75 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 49 โ”‚         243.50 / 245.17 ยฑ1.39 / 246.98 ms โ”‚        242.05 / 244.70 ยฑ2.27 / 248.12 ms โ”‚     no change โ”‚
   โ”‚ QQuery 50 โ”‚         182.51 / 188.03 ยฑ7.11 / 201.89 ms โ”‚        181.30 / 185.95 ยฑ3.88 / 192.72 ms โ”‚     no change โ”‚
   โ”‚ QQuery 51 โ”‚         170.66 / 174.10 ยฑ2.45 / 178.28 ms โ”‚        169.10 / 170.84 ยฑ1.48 / 172.62 ms โ”‚     no change โ”‚
   โ”‚ QQuery 52 โ”‚         107.15 / 108.15 ยฑ0.51 / 108.53 ms โ”‚        106.13 / 107.91 ยฑ1.35 / 110.20 ms โ”‚     no change โ”‚
   โ”‚ QQuery 53 โ”‚         103.03 / 105.07 ยฑ1.48 / 107.63 ms โ”‚        101.36 / 101.77 ยฑ0.29 / 102.24 ms โ”‚     no change โ”‚
   โ”‚ QQuery 54 โ”‚         143.93 / 145.92 ยฑ1.11 / 147.20 ms โ”‚        144.36 / 144.96 ยฑ0.57 / 145.99 ms โ”‚     no change โ”‚
   โ”‚ QQuery 55 โ”‚         105.78 / 107.68 ยฑ2.05 / 111.51 ms โ”‚        105.55 / 106.28 ยฑ0.78 / 107.51 ms โ”‚     no change โ”‚
   โ”‚ QQuery 56 โ”‚         138.45 / 139.67 ยฑ1.67 / 142.95 ms โ”‚        138.12 / 138.67 ยฑ0.46 / 139.49 ms โ”‚     no change โ”‚
   โ”‚ QQuery 57 โ”‚         163.66 / 165.43 ยฑ1.47 / 167.37 ms โ”‚        161.37 / 163.33 ยฑ1.66 / 165.70 ms โ”‚     no change โ”‚
   โ”‚ QQuery 58 โ”‚         246.06 / 251.06 ยฑ3.88 / 257.97 ms โ”‚        241.36 / 245.30 ยฑ2.93 / 249.75 ms โ”‚     no change โ”‚
   โ”‚ QQuery 59 โ”‚         200.92 / 203.48 ยฑ1.97 / 206.74 ms โ”‚        200.19 / 204.50 ยฑ3.71 / 210.20 ms โ”‚     no change โ”‚
   โ”‚ QQuery 60 โ”‚         144.86 / 147.25 ยฑ1.71 / 148.90 ms โ”‚        145.57 / 147.63 ยฑ2.14 / 151.72 ms โ”‚     no change โ”‚
   โ”‚ QQuery 61 โ”‚            14.27 / 15.56 ยฑ2.41 / 20.38 ms โ”‚           14.33 / 14.46 ยฑ0.07 / 14.56 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 62 โ”‚        855.51 / 874.12 ยฑ29.07 / 931.90 ms โ”‚       867.78 / 915.37 ยฑ38.10 / 957.00 ms โ”‚     no change โ”‚
   โ”‚ QQuery 63 โ”‚         102.47 / 103.15 ยฑ0.36 / 103.51 ms โ”‚        101.69 / 103.42 ยฑ2.15 / 107.54 ms โ”‚     no change โ”‚
   โ”‚ QQuery 64 โ”‚        628.45 / 641.05 ยฑ10.23 / 658.14 ms โ”‚        625.06 / 627.51 ยฑ2.13 / 630.45 ms โ”‚     no change โ”‚
   โ”‚ QQuery 65 โ”‚         240.95 / 243.82 ยฑ2.50 / 247.76 ms โ”‚        241.19 / 241.73 ยฑ0.42 / 242.28 ms โ”‚     no change โ”‚
   โ”‚ QQuery 66 โ”‚         212.92 / 219.65 ยฑ5.76 / 228.44 ms โ”‚        212.96 / 221.96 ยฑ9.87 / 236.50 ms โ”‚     no change โ”‚
   โ”‚ QQuery 67 โ”‚         289.12 / 294.94 ยฑ7.01 / 308.31 ms โ”‚        289.89 / 297.59 ยฑ5.89 / 307.11 ms โ”‚     no change โ”‚
   โ”‚ QQuery 68 โ”‚               8.82 / 9.03 ยฑ0.15 / 9.24 ms โ”‚              9.25 / 9.41 ยฑ0.14 / 9.64 ms โ”‚     no change โ”‚
   โ”‚ QQuery 69 โ”‚          98.79 / 101.72 ยฑ1.96 / 104.74 ms โ”‚         99.41 / 101.99 ยฑ1.52 / 103.65 ms โ”‚     no change โ”‚
   โ”‚ QQuery 70 โ”‚         323.07 / 331.08 ยฑ9.51 / 348.34 ms โ”‚       311.42 / 324.10 ยฑ10.28 / 339.21 ms โ”‚     no change โ”‚
   โ”‚ QQuery 71 โ”‚         137.75 / 141.32 ยฑ4.64 / 150.48 ms โ”‚        136.41 / 138.85 ยฑ2.58 / 143.56 ms โ”‚     no change โ”‚
   โ”‚ QQuery 72 โ”‚         538.00 / 547.15 ยฑ8.00 / 561.66 ms โ”‚       545.67 / 566.49 ยฑ21.72 / 600.16 ms โ”‚     no change โ”‚
   โ”‚ QQuery 73 โ”‚               6.56 / 6.70 ยฑ0.17 / 7.02 ms โ”‚              6.69 / 6.85 ยฑ0.20 / 7.23 ms โ”‚     no change โ”‚
   โ”‚ QQuery 74 โ”‚         546.56 / 558.76 ยฑ6.55 / 564.90 ms โ”‚       558.70 / 577.31 ยฑ22.75 / 619.57 ms โ”‚     no change โ”‚
   โ”‚ QQuery 75 โ”‚         262.98 / 265.71 ยฑ1.38 / 266.73 ms โ”‚        263.24 / 266.96 ยฑ3.53 / 273.59 ms โ”‚     no change โ”‚
   โ”‚ QQuery 76 โ”‚         127.52 / 130.10 ยฑ2.50 / 134.85 ms โ”‚        128.24 / 130.01 ยฑ1.94 / 133.33 ms โ”‚     no change โ”‚
   โ”‚ QQuery 77 โ”‚         186.03 / 187.72 ยฑ1.44 / 189.48 ms โ”‚        185.78 / 187.37 ยฑ0.80 / 187.82 ms โ”‚     no change โ”‚
   โ”‚ QQuery 78 โ”‚         305.73 / 307.45 ยฑ1.36 / 309.08 ms โ”‚        317.87 / 325.53 ยฑ5.04 / 333.04 ms โ”‚  1.06x slower โ”‚
   โ”‚ QQuery 79 โ”‚         227.07 / 230.72 ยฑ2.93 / 234.47 ms โ”‚        254.48 / 258.26 ยฑ3.25 / 263.19 ms โ”‚  1.12x slower โ”‚
   โ”‚ QQuery 80 โ”‚         299.69 / 302.83 ยฑ1.86 / 305.11 ms โ”‚        300.87 / 305.15 ยฑ2.55 / 308.55 ms โ”‚     no change โ”‚
   โ”‚ QQuery 81 โ”‚            27.09 / 27.31 ยฑ0.13 / 27.42 ms โ”‚           25.94 / 26.18 ยฑ0.14 / 26.35 ms โ”‚     no change โ”‚
   โ”‚ QQuery 82 โ”‚            40.55 / 40.84 ยฑ0.23 / 41.08 ms โ”‚           39.10 / 39.34 ยฑ0.17 / 39.54 ms โ”‚     no change โ”‚
   โ”‚ QQuery 83 โ”‚            36.82 / 37.22 ยฑ0.29 / 37.61 ms โ”‚           35.02 / 35.35 ยฑ0.23 / 35.72 ms โ”‚ +1.05x faster โ”‚
   โ”‚ QQuery 84 โ”‚            47.40 / 49.55 ยฑ2.72 / 54.85 ms โ”‚           46.02 / 47.53 ยฑ2.08 / 51.65 ms โ”‚     no change โ”‚
   โ”‚ QQuery 85 โ”‚         145.98 / 147.51 ยฑ1.73 / 150.83 ms โ”‚        143.03 / 145.04 ยฑ1.90 / 148.42 ms โ”‚     no change โ”‚
   โ”‚ QQuery 86 โ”‚            38.72 / 40.16 ยฑ2.25 / 44.62 ms โ”‚           37.41 / 37.72 ยฑ0.25 / 38.05 ms โ”‚ +1.06x faster โ”‚
   โ”‚ QQuery 87 โ”‚               3.85 / 3.95 ยฑ0.11 / 4.16 ms โ”‚              3.55 / 3.65 ยฑ0.14 / 3.92 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 88 โ”‚         103.61 / 104.11 ยฑ0.33 / 104.49 ms โ”‚          99.21 / 99.87 ยฑ0.55 / 100.54 ms โ”‚     no change โ”‚
   โ”‚ QQuery 89 โ”‚         119.84 / 122.91 ยฑ3.93 / 130.60 ms โ”‚        115.26 / 118.91 ยฑ6.19 / 131.27 ms โ”‚     no change โ”‚
   โ”‚ QQuery 90 โ”‚            23.14 / 23.88 ยฑ0.60 / 24.66 ms โ”‚           22.08 / 22.56 ยฑ0.49 / 23.42 ms โ”‚ +1.06x faster โ”‚
   โ”‚ QQuery 91 โ”‚            61.71 / 62.37 ยฑ0.61 / 63.47 ms โ”‚           59.18 / 59.66 ยฑ0.38 / 60.14 ms โ”‚     no change โ”‚
   โ”‚ QQuery 92 โ”‚            58.20 / 58.60 ยฑ0.24 / 58.88 ms โ”‚           56.29 / 56.67 ยฑ0.34 / 57.06 ms โ”‚     no change โ”‚
   โ”‚ QQuery 93 โ”‚         169.06 / 172.21 ยฑ1.91 / 174.50 ms โ”‚        158.97 / 160.35 ยฑ0.73 / 161.01 ms โ”‚ +1.07x faster โ”‚
   โ”‚ QQuery 94 โ”‚            60.25 / 61.05 ยฑ1.08 / 63.17 ms โ”‚           60.69 / 61.36 ยฑ0.41 / 61.85 ms โ”‚     no change โ”‚
   โ”‚ QQuery 95 โ”‚         111.44 / 111.81 ยฑ0.40 / 112.55 ms โ”‚        111.05 / 111.46 ยฑ0.34 / 111.80 ms โ”‚     no change โ”‚
   โ”‚ QQuery 96 โ”‚            67.16 / 67.67 ยฑ0.34 / 68.02 ms โ”‚           68.02 / 68.76 ยฑ0.47 / 69.23 ms โ”‚     no change โ”‚
   โ”‚ QQuery 97 โ”‚         111.40 / 115.14 ยฑ3.03 / 117.78 ms โ”‚        111.56 / 113.13 ยฑ1.78 / 116.43 ms โ”‚     no change โ”‚
   โ”‚ QQuery 98 โ”‚         148.07 / 151.05 ยฑ2.47 / 154.29 ms โ”‚        148.30 / 149.99 ยฑ2.05 / 154.03 ms โ”‚     no change โ”‚
   โ”‚ QQuery 99 โ”‚ 10703.28 / 10900.15 ยฑ138.22 / 11092.72 ms โ”‚ 10760.28 / 10891.23 ยฑ87.62 / 11000.91 ms โ”‚     no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 30469.30ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 30233.39ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚   307.77ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚   305.39ms โ”‚
   โ”‚ Queries Faster                                    โ”‚         12 โ”‚
   โ”‚ Queries Slower                                    โ”‚         12 โ”‚
   โ”‚ Queries with No Change                            โ”‚         75 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **tpcds โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 155.0s |
   | Peak memory | 6.3 GiB |
   | Avg memory | 5.6 GiB |
   | CPU user | 241.9s |
   | CPU sys | 8.0s |
   | Peak spill | 0 B |
   
   **tpcds โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 155.0s |
   | Peak memory | 6.4 GiB |
   | Avg memory | 5.7 GiB |
   | CPU user | 239.9s |
   | CPU sys | 7.7s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363696566

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4363668679)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark clickbench_partitioned.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                  HEAD โ”ƒ    speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 0  โ”‚          1.22 / 4.67 ยฑ6.82 / 18.31 ms โ”‚          1.19 / 4.58 ยฑ6.66 / 17.90 ms โ”‚     no change โ”‚
   โ”‚ QQuery 1  โ”‚        12.47 / 12.91 ยฑ0.23 / 13.11 ms โ”‚        12.20 / 12.47 ยฑ0.21 / 12.73 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚        37.64 / 37.99 ยฑ0.27 / 38.45 ms โ”‚        36.79 / 37.97 ยฑ1.17 / 39.86 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚        31.53 / 32.36 ยฑ0.57 / 33.23 ms โ”‚        31.66 / 31.94 ยฑ0.33 / 32.56 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚     233.75 / 236.79 ยฑ3.32 / 242.64 ms โ”‚     203.20 / 205.48 ยฑ2.00 / 207.82 ms โ”‚ +1.15x faster โ”‚
   โ”‚ QQuery 5  โ”‚     281.59 / 282.92 ยฑ1.17 / 284.93 ms โ”‚     277.57 / 280.25 ยฑ2.61 / 283.88 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚           6.90 / 7.71 ยฑ0.57 / 8.39 ms โ”‚           6.33 / 7.12 ยฑ0.59 / 8.08 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 7  โ”‚        14.04 / 14.13 ยฑ0.08 / 14.26 ms โ”‚        13.30 / 13.40 ยฑ0.08 / 13.47 ms โ”‚ +1.06x faster โ”‚
   โ”‚ QQuery 8  โ”‚     322.10 / 323.56 ยฑ1.57 / 326.51 ms โ”‚     312.27 / 314.02 ยฑ0.96 / 315.13 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚    456.60 / 464.77 ยฑ10.66 / 485.22 ms โ”‚     437.46 / 446.30 ยฑ6.76 / 457.53 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚        74.67 / 75.18 ยฑ0.39 / 75.85 ms โ”‚        72.00 / 72.70 ยฑ0.56 / 73.62 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚       85.46 / 90.13 ยฑ7.36 / 104.72 ms โ”‚        82.86 / 83.75 ยฑ0.60 / 84.56 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 12 โ”‚     272.81 / 277.50 ยฑ4.42 / 284.45 ms โ”‚     269.61 / 273.00 ยฑ2.88 / 276.74 ms โ”‚     no change โ”‚
   โ”‚ QQuery 13 โ”‚    382.48 / 391.82 ยฑ11.83 / 414.19 ms โ”‚    378.90 / 390.16 ยฑ11.21 / 411.09 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚     285.73 / 289.11 ยฑ4.58 / 297.74 ms โ”‚     279.69 / 283.28 ยฑ3.89 / 290.75 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚     272.03 / 279.25 ยฑ4.40 / 283.85 ms โ”‚     245.11 / 252.08 ยฑ4.87 / 259.70 ms โ”‚ +1.11x faster โ”‚
   โ”‚ QQuery 16 โ”‚     612.10 / 619.59 ยฑ5.50 / 625.54 ms โ”‚     606.07 / 610.26 ยฑ2.34 / 612.68 ms โ”‚     no change โ”‚
   โ”‚ QQuery 17 โ”‚     612.44 / 620.66 ยฑ4.99 / 627.12 ms โ”‚     608.55 / 615.04 ยฑ5.01 / 621.15 ms โ”‚     no change โ”‚
   โ”‚ QQuery 18 โ”‚  1238.59 / 1250.10 ยฑ7.34 / 1261.21 ms โ”‚ 1220.75 / 1236.67 ยฑ12.50 / 1255.87 ms โ”‚     no change โ”‚
   โ”‚ QQuery 19 โ”‚        28.93 / 34.99 ยฑ5.08 / 41.60 ms โ”‚        28.55 / 30.93 ยฑ3.67 / 38.07 ms โ”‚ +1.13x faster โ”‚
   โ”‚ QQuery 20 โ”‚     521.30 / 531.62 ยฑ9.48 / 546.54 ms โ”‚    522.67 / 534.06 ยฑ12.11 / 556.83 ms โ”‚     no change โ”‚
   โ”‚ QQuery 21 โ”‚     598.16 / 605.62 ยฑ7.77 / 618.73 ms โ”‚     595.16 / 599.27 ยฑ3.14 / 604.84 ms โ”‚     no change โ”‚
   โ”‚ QQuery 22 โ”‚  1065.43 / 1078.89 ยฑ9.11 / 1091.33 ms โ”‚ 1066.39 / 1077.66 ยฑ10.07 / 1094.13 ms โ”‚     no change โ”‚
   โ”‚ QQuery 23 โ”‚ 3360.63 / 3376.98 ยฑ10.18 / 3387.29 ms โ”‚ 3317.36 / 3339.06 ยฑ11.62 / 3349.71 ms โ”‚     no change โ”‚
   โ”‚ QQuery 24 โ”‚        41.83 / 44.52 ยฑ3.44 / 51.18 ms โ”‚        42.77 / 45.73 ยฑ2.83 / 50.17 ms โ”‚     no change โ”‚
   โ”‚ QQuery 25 โ”‚     114.49 / 117.86 ยฑ2.92 / 122.18 ms โ”‚     112.40 / 113.73 ยฑ1.13 / 115.61 ms โ”‚     no change โ”‚
   โ”‚ QQuery 26 โ”‚        42.56 / 43.11 ยฑ0.51 / 44.04 ms โ”‚        42.07 / 45.28 ยฑ3.24 / 50.16 ms โ”‚  1.05x slower โ”‚
   โ”‚ QQuery 27 โ”‚     674.39 / 678.50 ยฑ3.39 / 683.75 ms โ”‚     666.01 / 674.47 ยฑ5.46 / 681.81 ms โ”‚     no change โ”‚
   โ”‚ QQuery 28 โ”‚ 2995.62 / 3011.11 ยฑ16.60 / 3043.39 ms โ”‚ 3002.12 / 3029.85 ยฑ15.71 / 3045.23 ms โ”‚     no change โ”‚
   โ”‚ QQuery 29 โ”‚        42.70 / 47.53 ยฑ6.47 / 59.25 ms โ”‚        42.28 / 43.62 ยฑ2.02 / 47.63 ms โ”‚ +1.09x faster โ”‚
   โ”‚ QQuery 30 โ”‚     309.92 / 313.17 ยฑ3.75 / 320.30 ms โ”‚     305.48 / 308.21 ยฑ2.42 / 312.56 ms โ”‚     no change โ”‚
   โ”‚ QQuery 31 โ”‚     300.52 / 309.69 ยฑ6.39 / 317.19 ms โ”‚    291.06 / 304.59 ยฑ12.37 / 323.94 ms โ”‚     no change โ”‚
   โ”‚ QQuery 32 โ”‚   971.91 / 986.15 ยฑ12.44 / 1007.95 ms โ”‚     961.19 / 972.50 ยฑ8.12 / 985.83 ms โ”‚     no change โ”‚
   โ”‚ QQuery 33 โ”‚ 1407.36 / 1432.48 ยฑ13.61 / 1446.62 ms โ”‚  1423.65 / 1433.12 ยฑ7.75 / 1444.22 ms โ”‚     no change โ”‚
   โ”‚ QQuery 34 โ”‚ 1436.55 / 1456.88 ยฑ17.40 / 1482.24 ms โ”‚  1443.93 / 1451.47 ยฑ7.99 / 1466.14 ms โ”‚     no change โ”‚
   โ”‚ QQuery 35 โ”‚    280.17 / 315.28 ยฑ35.95 / 375.17 ms โ”‚    229.21 / 255.84 ยฑ39.82 / 334.33 ms โ”‚ +1.23x faster โ”‚
   โ”‚ QQuery 36 โ”‚        62.07 / 65.98 ยฑ4.39 / 73.42 ms โ”‚        63.91 / 70.16 ยฑ6.23 / 81.10 ms โ”‚  1.06x slower โ”‚
   โ”‚ QQuery 37 โ”‚       35.76 / 41.97 ยฑ10.97 / 63.88 ms โ”‚        36.35 / 40.06 ยฑ5.71 / 51.30 ms โ”‚     no change โ”‚
   โ”‚ QQuery 38 โ”‚        42.69 / 44.48 ยฑ1.34 / 46.76 ms โ”‚        40.29 / 41.88 ยฑ1.44 / 44.32 ms โ”‚ +1.06x faster โ”‚
   โ”‚ QQuery 39 โ”‚     125.52 / 132.99 ยฑ5.31 / 140.16 ms โ”‚     132.00 / 136.90 ยฑ4.71 / 144.94 ms โ”‚     no change โ”‚
   โ”‚ QQuery 40 โ”‚        14.83 / 17.57 ยฑ2.24 / 19.88 ms โ”‚        13.98 / 15.06 ยฑ1.16 / 17.24 ms โ”‚ +1.17x faster โ”‚
   โ”‚ QQuery 41 โ”‚        13.87 / 14.41 ยฑ0.37 / 14.91 ms โ”‚        13.60 / 13.71 ยฑ0.08 / 13.80 ms โ”‚     no change โ”‚
   โ”‚ QQuery 42 โ”‚        13.58 / 15.00 ยฑ2.40 / 19.80 ms โ”‚        12.74 / 13.00 ยฑ0.19 / 13.27 ms โ”‚ +1.15x faster โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 20027.96ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 19760.65ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚   465.77ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚   459.55ms โ”‚
   โ”‚ Queries Faster                                    โ”‚         11 โ”‚
   โ”‚ Queries Slower                                    โ”‚          2 โ”‚
   โ”‚ Queries with No Change                            โ”‚         30 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **clickbench_partitioned โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 105.0s |
   | Peak memory | 30.5 GiB |
   | Avg memory | 23.0 GiB |
   | CPU user | 1058.0s |
   | CPU sys | 63.8s |
   | Peak spill | 0 B |
   
   **clickbench_partitioned โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 100.0s |
   | Peak memory | 30.8 GiB |
   | Avg memory | 23.5 GiB |
   | CPU user | 1043.3s |
   | CPU sys | 63.6s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363624010

   run benchmarks


-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364091518

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364085635)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4364085635-1975-k9flr 6.12.68+ #1 SMP Wed Apr  1 02:23:28 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (08629a74dc3ee2842f44e43fe8ee96f8c3af7dea) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..08629a74dc3ee2842f44e43fe8ee96f8c3af7dea) using: clickbench_partitioned
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4360978892

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4360963663)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4360963663-1964-8kgp4 6.12.55+ #1 SMP Sun Feb  1 08:59:41 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (b79d7872b5b6070200a1b0838bf434ed6b3a6b87) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..b79d7872b5b6070200a1b0838bf434ed6b3a6b87) using: clickbench_partitioned
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4360976788

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4360963663)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4360963663-1966-5tpcs 6.12.55+ #1 SMP Sun Feb  1 08:59:41 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (b79d7872b5b6070200a1b0838bf434ed6b3a6b87) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..b79d7872b5b6070200a1b0838bf434ed6b3a6b87) using: tpch
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "comphead (via GitHub)" <gi...@apache.org>.
comphead commented on code in PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#discussion_r3176122449


##########
datafusion/physical-plan/benches/aggregate_vectorized.rs:
##########
@@ -179,6 +182,53 @@ fn primitive_vectorized_append(c: &mut Criterion) {
     group.finish();
 }
 
+fn single_group_by_primitive_intern(c: &mut Criterion) {
+    const BATCH_SIZE: usize = 4096;

Review Comment:
   should be batch size be like default? 



-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364584630

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364579262)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4364579262-1983-6tvb2 6.12.68+ #1 SMP Wed Apr  1 02:23:28 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (877ea557b9b6019f3544956a5d2ddd2c90921730) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..877ea557b9b6019f3544956a5d2ddd2c90921730) using: tpcds
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364116467

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364085635)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark tpch_sf1.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                           HEAD โ”ƒ speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 1  โ”‚ 40.04 / 41.48 ยฑ1.35 / 43.65 ms โ”‚     39.70 / 40.98 ยฑ1.41 / 42.95 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚ 21.22 / 21.32 ยฑ0.08 / 21.44 ms โ”‚     20.51 / 20.94 ยฑ0.32 / 21.48 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚ 35.74 / 38.05 ยฑ1.83 / 40.66 ms โ”‚     36.93 / 39.22 ยฑ1.73 / 41.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚ 18.26 / 19.52 ยฑ1.24 / 21.55 ms โ”‚     18.35 / 18.72 ยฑ0.63 / 19.98 ms โ”‚     no change โ”‚
   โ”‚ QQuery 5  โ”‚ 42.95 / 44.89 ยฑ1.47 / 46.88 ms โ”‚     43.55 / 45.36 ยฑ1.05 / 46.78 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚ 17.11 / 18.60 ยฑ1.02 / 19.93 ms โ”‚     16.87 / 17.28 ยฑ0.30 / 17.77 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 7  โ”‚ 49.95 / 51.90 ยฑ1.42 / 53.83 ms โ”‚     50.08 / 50.47 ยฑ0.30 / 50.74 ms โ”‚     no change โ”‚
   โ”‚ QQuery 8  โ”‚ 46.64 / 46.84 ยฑ0.12 / 47.00 ms โ”‚     46.55 / 47.43 ยฑ1.20 / 49.68 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚ 51.48 / 52.88 ยฑ1.24 / 54.87 ms โ”‚     51.46 / 52.68 ยฑ1.05 / 54.31 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚ 65.44 / 66.24 ยฑ1.20 / 68.62 ms โ”‚     64.77 / 65.07 ยฑ0.18 / 65.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚ 13.98 / 14.15 ยฑ0.13 / 14.33 ms โ”‚     13.96 / 14.53 ยฑ0.57 / 15.44 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚ 26.46 / 26.64 ยฑ0.12 / 26.84 ms โ”‚     26.21 / 26.59 ยฑ0.38 / 27.16 ms โ”‚     no change โ”‚
   โ”‚ QQuery 13 โ”‚ 35.34 / 36.23 ยฑ0.53 / 36.96 ms โ”‚     36.07 / 36.85 ยฑ0.52 / 37.45 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚ 26.36 / 26.66 ยฑ0.17 / 26.87 ms โ”‚     26.48 / 26.51 ยฑ0.03 / 26.56 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚ 33.00 / 33.49 ยฑ0.70 / 34.87 ms โ”‚     32.81 / 33.45 ยฑ0.70 / 34.78 ms โ”‚     no change โ”‚
   โ”‚ QQuery 16 โ”‚ 15.20 / 15.40 ยฑ0.15 / 15.65 ms โ”‚     15.29 / 15.36 ยฑ0.08 / 15.49 ms โ”‚     no change โ”‚
   โ”‚ QQuery 17 โ”‚ 76.86 / 77.69 ยฑ1.00 / 79.42 ms โ”‚     74.26 / 75.60 ยฑ1.10 / 77.27 ms โ”‚     no change โ”‚
   โ”‚ QQuery 18 โ”‚ 69.31 / 69.68 ยฑ0.31 / 70.18 ms โ”‚     67.64 / 69.12 ยฑ0.82 / 69.94 ms โ”‚     no change โ”‚
   โ”‚ QQuery 19 โ”‚ 37.47 / 37.78 ยฑ0.23 / 38.05 ms โ”‚     37.19 / 37.66 ยฑ0.58 / 38.77 ms โ”‚     no change โ”‚
   โ”‚ QQuery 20 โ”‚ 38.95 / 39.16 ยฑ0.24 / 39.59 ms โ”‚     38.86 / 39.03 ยฑ0.16 / 39.32 ms โ”‚     no change โ”‚
   โ”‚ QQuery 21 โ”‚ 60.81 / 61.76 ยฑ0.92 / 63.42 ms โ”‚     59.20 / 60.62 ยฑ1.04 / 62.18 ms โ”‚     no change โ”‚
   โ”‚ QQuery 22 โ”‚ 23.78 / 24.09 ยฑ0.25 / 24.48 ms โ”‚     23.66 / 23.90 ยฑ0.26 / 24.27 ms โ”‚     no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ          โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 864.43ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 857.38ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚  39.29ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚  38.97ms โ”‚
   โ”‚ Queries Faster                                    โ”‚        1 โ”‚
   โ”‚ Queries Slower                                    โ”‚        0 โ”‚
   โ”‚ Queries with No Change                            โ”‚       21 โ”‚
   โ”‚ Queries with Failure                              โ”‚        0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **tpch โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 5.0s |
   | Peak memory | 5.5 GiB |
   | Avg memory | 5.0 GiB |
   | CPU user | 32.4s |
   | CPU sys | 2.0s |
   | Peak spill | 0 B |
   
   **tpch โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 5.0s |
   | Peak memory | 5.4 GiB |
   | Avg memory | 5.0 GiB |
   | CPU user | 32.0s |
   | CPU sys | 2.1s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4361048511

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4360963663)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark tpch_sf1.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                           HEAD โ”ƒ speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 1  โ”‚ 40.62 / 42.00 ยฑ1.12 / 43.28 ms โ”‚     40.09 / 41.48 ยฑ2.51 / 46.50 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚ 21.30 / 21.71 ยฑ0.60 / 22.91 ms โ”‚     21.35 / 21.68 ยฑ0.31 / 22.13 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚ 35.86 / 38.73 ยฑ1.45 / 39.73 ms โ”‚     36.22 / 37.90 ยฑ1.38 / 39.18 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚ 18.48 / 18.95 ยฑ0.70 / 20.33 ms โ”‚     18.33 / 18.84 ยฑ0.59 / 19.93 ms โ”‚     no change โ”‚
   โ”‚ QQuery 5  โ”‚ 44.53 / 46.49 ยฑ1.09 / 47.52 ms โ”‚     45.76 / 46.45 ยฑ0.41 / 47.04 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚ 17.31 / 17.65 ยฑ0.39 / 18.20 ms โ”‚     17.17 / 17.33 ยฑ0.13 / 17.48 ms โ”‚     no change โ”‚
   โ”‚ QQuery 7  โ”‚ 50.90 / 52.44 ยฑ1.16 / 54.08 ms โ”‚     49.87 / 51.02 ยฑ1.15 / 53.11 ms โ”‚     no change โ”‚
   โ”‚ QQuery 8  โ”‚ 47.31 / 47.70 ยฑ0.39 / 48.43 ms โ”‚     47.08 / 47.42 ยฑ0.43 / 48.25 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚ 51.56 / 53.02 ยฑ0.94 / 54.50 ms โ”‚     52.11 / 52.66 ยฑ0.30 / 52.90 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚ 65.91 / 67.20 ยฑ1.04 / 68.87 ms โ”‚     65.68 / 66.04 ยฑ0.27 / 66.38 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚ 14.06 / 14.26 ยฑ0.16 / 14.44 ms โ”‚     13.95 / 14.12 ยฑ0.19 / 14.50 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚ 26.61 / 26.83 ยฑ0.18 / 27.02 ms โ”‚     26.49 / 26.77 ยฑ0.23 / 27.15 ms โ”‚     no change โ”‚
   โ”‚ QQuery 13 โ”‚ 35.67 / 36.81 ยฑ0.64 / 37.58 ms โ”‚     35.31 / 36.03 ยฑ0.40 / 36.48 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚ 26.85 / 27.21 ยฑ0.21 / 27.49 ms โ”‚     26.78 / 27.00 ยฑ0.22 / 27.40 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚ 33.36 / 33.58 ยฑ0.20 / 33.83 ms โ”‚     32.96 / 33.25 ยฑ0.19 / 33.57 ms โ”‚     no change โ”‚
   โ”‚ QQuery 16 โ”‚ 15.36 / 15.54 ยฑ0.15 / 15.68 ms โ”‚     15.37 / 15.54 ยฑ0.09 / 15.61 ms โ”‚     no change โ”‚
   โ”‚ QQuery 17 โ”‚ 77.84 / 78.51 ยฑ0.50 / 79.38 ms โ”‚     70.91 / 72.12 ยฑ1.04 / 73.40 ms โ”‚ +1.09x faster โ”‚
   โ”‚ QQuery 18 โ”‚ 69.21 / 70.21 ยฑ0.75 / 71.36 ms โ”‚     68.84 / 70.23 ยฑ1.05 / 71.49 ms โ”‚     no change โ”‚
   โ”‚ QQuery 19 โ”‚ 37.89 / 38.28 ยฑ0.33 / 38.65 ms โ”‚     37.83 / 38.42 ยฑ0.52 / 39.18 ms โ”‚     no change โ”‚
   โ”‚ QQuery 20 โ”‚ 39.34 / 39.96 ยฑ0.54 / 40.74 ms โ”‚     39.02 / 39.37 ยฑ0.28 / 39.82 ms โ”‚     no change โ”‚
   โ”‚ QQuery 21 โ”‚ 60.67 / 62.47 ยฑ1.96 / 66.13 ms โ”‚     60.76 / 62.30 ยฑ1.27 / 64.47 ms โ”‚     no change โ”‚
   โ”‚ QQuery 22 โ”‚ 24.20 / 24.51 ยฑ0.31 / 25.07 ms โ”‚     24.24 / 24.42 ยฑ0.19 / 24.75 ms โ”‚     no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ          โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 874.07ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 860.39ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚  39.73ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚  39.11ms โ”‚
   โ”‚ Queries Faster                                    โ”‚        1 โ”‚
   โ”‚ Queries Slower                                    โ”‚        0 โ”‚
   โ”‚ Queries with No Change                            โ”‚       21 โ”‚
   โ”‚ Queries with Failure                              โ”‚        0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **tpch โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 5.0s |
   | Peak memory | 5.6 GiB |
   | Avg memory | 5.0 GiB |
   | CPU user | 32.5s |
   | CPU sys | 2.3s |
   | Peak spill | 0 B |
   
   **tpch โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 5.0s |
   | Peak memory | 5.5 GiB |
   | Avg memory | 5.0 GiB |
   | CPU user | 31.9s |
   | CPU sys | 2.4s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364584554

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364579262)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4364579262-1982-2wfps 6.12.68+ #1 SMP Wed Apr  1 02:23:28 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (877ea557b9b6019f3544956a5d2ddd2c90921730) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..877ea557b9b6019f3544956a5d2ddd2c90921730) using: clickbench_partitioned
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4365824085

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4365794880)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark tpch_sf10.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                               HEAD โ”ƒ speed-up-single-primitive-group-by โ”ƒ    Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 1  โ”‚  329.46 / 330.50 ยฑ1.29 / 332.92 ms โ”‚  324.70 / 326.95 ยฑ1.62 / 329.04 ms โ”‚ no change โ”‚
   โ”‚ QQuery 2  โ”‚  125.49 / 127.37 ยฑ1.36 / 129.65 ms โ”‚  123.97 / 126.93 ยฑ2.36 / 130.42 ms โ”‚ no change โ”‚
   โ”‚ QQuery 3  โ”‚  269.95 / 273.75 ยฑ3.92 / 278.66 ms โ”‚  266.78 / 269.18 ยฑ2.54 / 274.08 ms โ”‚ no change โ”‚
   โ”‚ QQuery 4  โ”‚  134.76 / 135.68 ยฑ0.55 / 136.49 ms โ”‚  133.38 / 134.86 ยฑ1.75 / 137.98 ms โ”‚ no change โ”‚
   โ”‚ QQuery 5  โ”‚  394.01 / 398.58 ยฑ5.25 / 408.68 ms โ”‚  387.93 / 394.39 ยฑ5.04 / 402.16 ms โ”‚ no change โ”‚
   โ”‚ QQuery 6  โ”‚  132.29 / 134.28 ยฑ2.05 / 137.32 ms โ”‚  130.90 / 131.95 ยฑ0.85 / 133.23 ms โ”‚ no change โ”‚
   โ”‚ QQuery 7  โ”‚  506.53 / 513.60 ยฑ4.83 / 520.62 ms โ”‚  517.12 / 523.27 ยฑ4.87 / 530.48 ms โ”‚ no change โ”‚
   โ”‚ QQuery 8  โ”‚  424.20 / 436.97 ยฑ6.89 / 443.69 ms โ”‚ 433.30 / 441.10 ยฑ11.80 / 464.49 ms โ”‚ no change โ”‚
   โ”‚ QQuery 9  โ”‚  605.60 / 620.34 ยฑ9.73 / 631.39 ms โ”‚ 604.45 / 617.54 ยฑ15.35 / 645.90 ms โ”‚ no change โ”‚
   โ”‚ QQuery 10 โ”‚  321.98 / 330.44 ยฑ9.39 / 348.01 ms โ”‚  317.92 / 326.39 ยฑ5.37 / 334.28 ms โ”‚ no change โ”‚
   โ”‚ QQuery 11 โ”‚   97.40 / 101.30 ยฑ5.32 / 111.69 ms โ”‚     95.01 / 96.29 ยฑ1.15 / 98.32 ms โ”‚ no change โ”‚
   โ”‚ QQuery 12 โ”‚  192.03 / 197.80 ยฑ4.68 / 205.54 ms โ”‚  195.51 / 199.41 ยฑ6.42 / 212.21 ms โ”‚ no change โ”‚
   โ”‚ QQuery 13 โ”‚  296.20 / 305.46 ยฑ7.66 / 319.37 ms โ”‚  288.60 / 297.15 ยฑ5.08 / 303.40 ms โ”‚ no change โ”‚
   โ”‚ QQuery 14 โ”‚  181.63 / 186.58 ยฑ5.43 / 196.87 ms โ”‚  181.72 / 183.70 ยฑ2.44 / 188.39 ms โ”‚ no change โ”‚
   โ”‚ QQuery 15 โ”‚  330.31 / 331.81 ยฑ1.73 / 334.86 ms โ”‚  319.35 / 321.41 ยฑ2.18 / 325.57 ms โ”‚ no change โ”‚
   โ”‚ QQuery 16 โ”‚     74.56 / 76.84 ยฑ2.63 / 81.86 ms โ”‚     73.46 / 76.08 ยฑ2.18 / 79.92 ms โ”‚ no change โ”‚
   โ”‚ QQuery 17 โ”‚  699.14 / 704.67 ยฑ3.62 / 709.33 ms โ”‚ 665.03 / 674.94 ยฑ13.33 / 700.50 ms โ”‚ no change โ”‚
   โ”‚ QQuery 18 โ”‚ 720.52 / 739.09 ยฑ13.84 / 756.66 ms โ”‚  728.45 / 740.15 ยฑ8.63 / 749.49 ms โ”‚ no change โ”‚
   โ”‚ QQuery 19 โ”‚ 271.75 / 283.57 ยฑ12.95 / 299.89 ms โ”‚ 268.28 / 284.97 ยฑ28.47 / 341.62 ms โ”‚ no change โ”‚
   โ”‚ QQuery 20 โ”‚  300.77 / 308.87 ยฑ9.20 / 326.87 ms โ”‚ 296.32 / 308.92 ยฑ10.11 / 322.35 ms โ”‚ no change โ”‚
   โ”‚ QQuery 21 โ”‚  743.66 / 751.10 ยฑ7.52 / 764.50 ms โ”‚  741.18 / 748.10 ยฑ5.29 / 756.77 ms โ”‚ no change โ”‚
   โ”‚ QQuery 22 โ”‚     65.00 / 68.38 ยฑ2.50 / 71.49 ms โ”‚     65.08 / 67.55 ยฑ1.74 / 69.43 ms โ”‚ no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ           โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 7356.97ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 7291.23ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚  334.41ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚  331.42ms โ”‚
   โ”‚ Queries Faster                                    โ”‚         0 โ”‚
   โ”‚ Queries Slower                                    โ”‚         0 โ”‚
   โ”‚ Queries with No Change                            โ”‚        22 โ”‚
   โ”‚ Queries with Failure                              โ”‚         0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **tpch10 โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 40.0s |
   | Peak memory | 12.3 GiB |
   | Avg memory | 8.4 GiB |
   | CPU user | 394.4s |
   | CPU sys | 19.8s |
   | Peak spill | 0 B |
   
   **tpch10 โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 40.0s |
   | Peak memory | 11.9 GiB |
   | Avg memory | 8.4 GiB |
   | CPU user | 389.7s |
   | CPU sys | 19.6s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364124746

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364085635)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark tpcds_sf1.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                     HEAD โ”ƒ       speed-up-single-primitive-group-by โ”ƒ       Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 1  โ”‚              6.86 / 7.36 ยฑ0.92 / 9.20 ms โ”‚              6.98 / 7.44 ยฑ0.82 / 9.07 ms โ”‚    no change โ”‚
   โ”‚ QQuery 2  โ”‚        144.33 / 144.95 ยฑ0.48 / 145.81 ms โ”‚        146.90 / 147.27 ยฑ0.38 / 147.93 ms โ”‚    no change โ”‚
   โ”‚ QQuery 3  โ”‚        113.22 / 115.61 ยฑ2.83 / 121.16 ms โ”‚        113.94 / 114.30 ยฑ0.28 / 114.65 ms โ”‚    no change โ”‚
   โ”‚ QQuery 4  โ”‚     1261.94 / 1271.87 ยฑ6.70 / 1281.16 ms โ”‚     1251.68 / 1264.62 ยฑ9.25 / 1276.16 ms โ”‚    no change โ”‚
   โ”‚ QQuery 5  โ”‚        173.48 / 175.31 ยฑ1.31 / 176.88 ms โ”‚        173.76 / 175.12 ยฑ1.57 / 178.03 ms โ”‚    no change โ”‚
   โ”‚ QQuery 6  โ”‚        129.33 / 131.19 ยฑ1.34 / 132.98 ms โ”‚        131.44 / 132.35 ยฑ0.71 / 133.58 ms โ”‚    no change โ”‚
   โ”‚ QQuery 7  โ”‚        316.98 / 320.04 ยฑ1.66 / 321.62 ms โ”‚        318.61 / 320.00 ยฑ1.05 / 321.34 ms โ”‚    no change โ”‚
   โ”‚ QQuery 8  โ”‚        112.87 / 113.71 ยฑ0.91 / 115.40 ms โ”‚        113.11 / 113.82 ยฑ0.78 / 115.06 ms โ”‚    no change โ”‚
   โ”‚ QQuery 9  โ”‚          95.65 / 98.93 ยฑ2.17 / 100.89 ms โ”‚          96.57 / 99.69 ยฑ2.35 / 102.22 ms โ”‚    no change โ”‚
   โ”‚ QQuery 10 โ”‚        102.69 / 103.97 ยฑ1.76 / 107.43 ms โ”‚        100.80 / 102.90 ยฑ1.78 / 105.95 ms โ”‚    no change โ”‚
   โ”‚ QQuery 11 โ”‚       851.22 / 866.41 ยฑ10.19 / 877.88 ms โ”‚       854.28 / 874.08 ยฑ13.54 / 894.22 ms โ”‚    no change โ”‚
   โ”‚ QQuery 12 โ”‚           43.08 / 43.44 ยฑ0.25 / 43.81 ms โ”‚           42.91 / 43.04 ยฑ0.12 / 43.24 ms โ”‚    no change โ”‚
   โ”‚ QQuery 13 โ”‚        390.05 / 390.87 ยฑ0.99 / 392.76 ms โ”‚        387.62 / 389.98 ยฑ1.96 / 393.42 ms โ”‚    no change โ”‚
   โ”‚ QQuery 14 โ”‚     1027.41 / 1036.23 ยฑ6.77 / 1047.70 ms โ”‚     1029.26 / 1035.68 ยฑ4.48 / 1043.01 ms โ”‚    no change โ”‚
   โ”‚ QQuery 15 โ”‚           14.68 / 15.00 ยฑ0.22 / 15.29 ms โ”‚           14.58 / 15.83 ยฑ2.02 / 19.86 ms โ”‚ 1.06x slower โ”‚
   โ”‚ QQuery 16 โ”‚              7.36 / 7.46 ยฑ0.11 / 7.68 ms โ”‚              7.32 / 7.47 ยฑ0.15 / 7.76 ms โ”‚    no change โ”‚
   โ”‚ QQuery 17 โ”‚        203.42 / 204.23 ยฑ0.75 / 205.22 ms โ”‚        200.41 / 202.57 ยฑ1.25 / 204.08 ms โ”‚    no change โ”‚
   โ”‚ QQuery 18 โ”‚        123.87 / 124.49 ยฑ0.36 / 124.93 ms โ”‚        121.56 / 123.61 ยฑ1.11 / 124.89 ms โ”‚    no change โ”‚
   โ”‚ QQuery 19 โ”‚        153.50 / 154.60 ยฑ1.04 / 156.36 ms โ”‚        154.37 / 155.49 ยฑ0.88 / 157.03 ms โ”‚    no change โ”‚
   โ”‚ QQuery 20 โ”‚           13.31 / 13.44 ยฑ0.09 / 13.57 ms โ”‚           13.03 / 13.30 ยฑ0.19 / 13.62 ms โ”‚    no change โ”‚
   โ”‚ QQuery 21 โ”‚           18.90 / 19.14 ยฑ0.20 / 19.47 ms โ”‚           18.65 / 19.04 ยฑ0.21 / 19.26 ms โ”‚    no change โ”‚
   โ”‚ QQuery 22 โ”‚        474.39 / 478.30 ยฑ2.72 / 482.86 ms โ”‚        474.25 / 477.72 ยฑ2.38 / 481.27 ms โ”‚    no change โ”‚
   โ”‚ QQuery 23 โ”‚     1009.42 / 1014.88 ยฑ3.68 / 1019.34 ms โ”‚     1007.38 / 1016.09 ยฑ7.59 / 1029.96 ms โ”‚    no change โ”‚
   โ”‚ QQuery 24 โ”‚        634.37 / 638.43 ยฑ3.94 / 645.36 ms โ”‚        624.35 / 633.89 ยฑ5.41 / 640.68 ms โ”‚    no change โ”‚
   โ”‚ QQuery 25 โ”‚        304.93 / 306.85 ยฑ1.54 / 308.37 ms โ”‚        303.02 / 305.03 ยฑ1.09 / 306.25 ms โ”‚    no change โ”‚
   โ”‚ QQuery 26 โ”‚           76.18 / 76.93 ยฑ0.48 / 77.58 ms โ”‚           75.39 / 76.12 ยฑ0.85 / 77.24 ms โ”‚    no change โ”‚
   โ”‚ QQuery 27 โ”‚              7.02 / 7.14 ยฑ0.18 / 7.49 ms โ”‚              6.95 / 7.09 ยฑ0.13 / 7.32 ms โ”‚    no change โ”‚
   โ”‚ QQuery 28 โ”‚        147.80 / 149.25 ยฑ1.18 / 151.40 ms โ”‚        147.09 / 148.07 ยฑ0.56 / 148.67 ms โ”‚    no change โ”‚
   โ”‚ QQuery 29 โ”‚        244.07 / 247.65 ยฑ6.18 / 260.00 ms โ”‚        244.22 / 248.95 ยฑ5.45 / 258.43 ms โ”‚    no change โ”‚
   โ”‚ QQuery 30 โ”‚           41.05 / 41.79 ยฑ0.73 / 43.03 ms โ”‚           41.26 / 41.64 ยฑ0.35 / 42.22 ms โ”‚    no change โ”‚
   โ”‚ QQuery 31 โ”‚        164.81 / 165.73 ยฑ0.67 / 166.54 ms โ”‚        164.97 / 166.99 ยฑ2.57 / 171.99 ms โ”‚    no change โ”‚
   โ”‚ QQuery 32 โ”‚           13.23 / 13.44 ยฑ0.25 / 13.90 ms โ”‚           13.23 / 13.43 ยฑ0.18 / 13.75 ms โ”‚    no change โ”‚
   โ”‚ QQuery 33 โ”‚        138.36 / 139.43 ยฑ1.24 / 141.81 ms โ”‚        138.58 / 139.48 ยฑ0.65 / 140.50 ms โ”‚    no change โ”‚
   โ”‚ QQuery 34 โ”‚              7.15 / 7.29 ยฑ0.15 / 7.59 ms โ”‚             6.97 / 7.83 ยฑ1.44 / 10.67 ms โ”‚ 1.07x slower โ”‚
   โ”‚ QQuery 35 โ”‚        100.01 / 101.03 ยฑ0.65 / 101.99 ms โ”‚         99.75 / 100.85 ยฑ0.68 / 101.57 ms โ”‚    no change โ”‚
   โ”‚ QQuery 36 โ”‚              6.62 / 6.95 ยฑ0.21 / 7.18 ms โ”‚              6.69 / 6.88 ยฑ0.13 / 7.04 ms โ”‚    no change โ”‚
   โ”‚ QQuery 37 โ”‚              8.23 / 8.33 ยฑ0.07 / 8.42 ms โ”‚              8.21 / 8.34 ยฑ0.10 / 8.44 ms โ”‚    no change โ”‚
   โ”‚ QQuery 38 โ”‚           85.15 / 87.58 ยฑ3.42 / 94.37 ms โ”‚           84.63 / 86.85 ยฑ1.54 / 88.63 ms โ”‚    no change โ”‚
   โ”‚ QQuery 39 โ”‚        115.72 / 116.86 ยฑ1.54 / 119.85 ms โ”‚        116.44 / 117.42 ยฑ1.02 / 118.84 ms โ”‚    no change โ”‚
   โ”‚ QQuery 40 โ”‚           93.23 / 95.97 ยฑ2.24 / 99.72 ms โ”‚           91.74 / 94.86 ยฑ2.57 / 98.66 ms โ”‚    no change โ”‚
   โ”‚ QQuery 41 โ”‚           14.55 / 14.78 ยฑ0.23 / 15.20 ms โ”‚           14.41 / 14.55 ยฑ0.23 / 15.00 ms โ”‚    no change โ”‚
   โ”‚ QQuery 42 โ”‚        106.91 / 107.34 ยฑ0.29 / 107.67 ms โ”‚        106.43 / 107.76 ยฑ0.68 / 108.32 ms โ”‚    no change โ”‚
   โ”‚ QQuery 43 โ”‚              5.90 / 6.60 ยฑ1.24 / 9.08 ms โ”‚              5.76 / 6.48 ยฑ1.26 / 9.00 ms โ”‚    no change โ”‚
   โ”‚ QQuery 44 โ”‚           11.01 / 11.10 ยฑ0.07 / 11.22 ms โ”‚           10.70 / 10.86 ยฑ0.15 / 11.10 ms โ”‚    no change โ”‚
   โ”‚ QQuery 45 โ”‚           44.53 / 45.28 ยฑ0.99 / 47.16 ms โ”‚           43.33 / 44.08 ยฑ0.49 / 44.62 ms โ”‚    no change โ”‚
   โ”‚ QQuery 46 โ”‚              8.51 / 8.76 ยฑ0.25 / 9.23 ms โ”‚             8.42 / 9.28 ยฑ1.53 / 12.33 ms โ”‚ 1.06x slower โ”‚
   โ”‚ QQuery 47 โ”‚        683.09 / 691.08 ยฑ5.76 / 700.17 ms โ”‚        673.10 / 680.10 ยฑ5.55 / 688.65 ms โ”‚    no change โ”‚
   โ”‚ QQuery 48 โ”‚        273.25 / 277.79 ยฑ7.35 / 292.44 ms โ”‚        271.16 / 276.48 ยฑ3.27 / 281.08 ms โ”‚    no change โ”‚
   โ”‚ QQuery 49 โ”‚        243.64 / 245.01 ยฑ0.87 / 246.37 ms โ”‚        243.52 / 245.61 ยฑ1.08 / 246.57 ms โ”‚    no change โ”‚
   โ”‚ QQuery 50 โ”‚        180.57 / 185.57 ยฑ3.35 / 190.42 ms โ”‚        181.18 / 185.76 ยฑ3.66 / 189.68 ms โ”‚    no change โ”‚
   โ”‚ QQuery 51 โ”‚        171.92 / 173.58 ยฑ1.60 / 176.52 ms โ”‚        170.67 / 174.12 ยฑ2.16 / 176.85 ms โ”‚    no change โ”‚
   โ”‚ QQuery 52 โ”‚        107.30 / 109.38 ยฑ2.57 / 114.44 ms โ”‚        108.11 / 110.29 ยฑ2.88 / 115.90 ms โ”‚    no change โ”‚
   โ”‚ QQuery 53 โ”‚        101.59 / 102.50 ยฑ0.52 / 103.08 ms โ”‚        102.87 / 103.87 ยฑ0.89 / 105.30 ms โ”‚    no change โ”‚
   โ”‚ QQuery 54 โ”‚        144.42 / 147.54 ยฑ2.72 / 152.44 ms โ”‚        145.75 / 147.87 ยฑ1.92 / 150.92 ms โ”‚    no change โ”‚
   โ”‚ QQuery 55 โ”‚        106.73 / 107.88 ยฑ0.97 / 108.99 ms โ”‚        106.89 / 108.05 ยฑ1.22 / 110.25 ms โ”‚    no change โ”‚
   โ”‚ QQuery 56 โ”‚        138.63 / 139.50 ยฑ0.62 / 140.14 ms โ”‚        139.41 / 140.88 ยฑ0.96 / 142.13 ms โ”‚    no change โ”‚
   โ”‚ QQuery 57 โ”‚        163.49 / 165.29 ยฑ1.47 / 167.88 ms โ”‚        163.02 / 164.82 ยฑ1.79 / 167.41 ms โ”‚    no change โ”‚
   โ”‚ QQuery 58 โ”‚        242.19 / 243.48 ยฑ0.69 / 244.22 ms โ”‚        242.94 / 244.19 ยฑ0.88 / 245.64 ms โ”‚    no change โ”‚
   โ”‚ QQuery 59 โ”‚        192.21 / 194.05 ยฑ1.58 / 196.52 ms โ”‚        192.84 / 195.57 ยฑ1.50 / 196.98 ms โ”‚    no change โ”‚
   โ”‚ QQuery 60 โ”‚        140.52 / 142.18 ยฑ1.60 / 144.90 ms โ”‚        140.62 / 142.41 ยฑ1.81 / 145.89 ms โ”‚    no change โ”‚
   โ”‚ QQuery 61 โ”‚           13.34 / 13.55 ยฑ0.23 / 13.98 ms โ”‚           13.46 / 13.56 ยฑ0.14 / 13.83 ms โ”‚    no change โ”‚
   โ”‚ QQuery 62 โ”‚        855.48 / 858.69 ยฑ3.64 / 865.76 ms โ”‚        852.75 / 858.04 ยฑ4.23 / 865.11 ms โ”‚    no change โ”‚
   โ”‚ QQuery 63 โ”‚        102.23 / 103.99 ยฑ1.77 / 106.89 ms โ”‚        102.65 / 104.26 ยฑ2.00 / 108.20 ms โ”‚    no change โ”‚
   โ”‚ QQuery 64 โ”‚        627.08 / 633.29 ยฑ4.05 / 639.61 ms โ”‚        625.11 / 629.47 ยฑ3.71 / 634.75 ms โ”‚    no change โ”‚
   โ”‚ QQuery 65 โ”‚        240.82 / 244.42 ยฑ4.37 / 252.70 ms โ”‚        242.18 / 244.18 ยฑ1.36 / 246.28 ms โ”‚    no change โ”‚
   โ”‚ QQuery 66 โ”‚        213.56 / 219.62 ยฑ5.18 / 227.56 ms โ”‚        213.50 / 220.25 ยฑ7.59 / 233.15 ms โ”‚    no change โ”‚
   โ”‚ QQuery 67 โ”‚        289.29 / 297.61 ยฑ8.57 / 313.21 ms โ”‚        290.68 / 295.63 ยฑ4.10 / 300.18 ms โ”‚    no change โ”‚
   โ”‚ QQuery 68 โ”‚              8.78 / 8.99 ยฑ0.18 / 9.32 ms โ”‚              8.69 / 8.82 ยฑ0.21 / 9.24 ms โ”‚    no change โ”‚
   โ”‚ QQuery 69 โ”‚           98.28 / 98.53 ยฑ0.21 / 98.83 ms โ”‚           96.52 / 96.91 ยฑ0.32 / 97.37 ms โ”‚    no change โ”‚
   โ”‚ QQuery 70 โ”‚        314.17 / 322.98 ยฑ8.45 / 338.72 ms โ”‚        309.28 / 319.00 ยฑ5.62 / 326.74 ms โ”‚    no change โ”‚
   โ”‚ QQuery 71 โ”‚        132.49 / 134.03 ยฑ0.81 / 134.81 ms โ”‚        132.94 / 134.76 ยฑ1.58 / 137.60 ms โ”‚    no change โ”‚
   โ”‚ QQuery 72 โ”‚       532.17 / 545.62 ยฑ11.78 / 561.11 ms โ”‚        538.26 / 542.98 ยฑ3.65 / 547.42 ms โ”‚    no change โ”‚
   โ”‚ QQuery 73 โ”‚              6.60 / 6.78 ยฑ0.20 / 7.17 ms โ”‚              6.74 / 6.89 ยฑ0.22 / 7.33 ms โ”‚    no change โ”‚
   โ”‚ QQuery 74 โ”‚        548.38 / 554.38 ยฑ4.40 / 561.00 ms โ”‚        537.39 / 541.70 ยฑ3.52 / 545.93 ms โ”‚    no change โ”‚
   โ”‚ QQuery 75 โ”‚        265.77 / 266.29 ยฑ0.52 / 267.09 ms โ”‚        265.10 / 267.00 ยฑ2.08 / 270.89 ms โ”‚    no change โ”‚
   โ”‚ QQuery 76 โ”‚        130.72 / 131.86 ยฑ1.92 / 135.67 ms โ”‚        129.15 / 131.26 ยฑ2.87 / 136.93 ms โ”‚    no change โ”‚
   โ”‚ QQuery 77 โ”‚        187.39 / 188.21 ยฑ0.86 / 189.74 ms โ”‚        186.61 / 188.33 ยฑ1.94 / 191.89 ms โ”‚    no change โ”‚
   โ”‚ QQuery 78 โ”‚        308.11 / 310.76 ยฑ2.17 / 314.31 ms โ”‚        306.86 / 308.90 ยฑ1.87 / 312.32 ms โ”‚    no change โ”‚
   โ”‚ QQuery 79 โ”‚        227.01 / 228.40 ยฑ1.13 / 229.86 ms โ”‚        229.63 / 230.93 ยฑ1.40 / 232.94 ms โ”‚    no change โ”‚
   โ”‚ QQuery 80 โ”‚        301.52 / 302.60 ยฑ1.19 / 304.81 ms โ”‚        299.86 / 302.09 ยฑ2.13 / 305.19 ms โ”‚    no change โ”‚
   โ”‚ QQuery 81 โ”‚           25.57 / 26.36 ยฑ0.44 / 26.90 ms โ”‚           25.47 / 25.78 ยฑ0.24 / 26.11 ms โ”‚    no change โ”‚
   โ”‚ QQuery 82 โ”‚           38.96 / 39.21 ยฑ0.17 / 39.39 ms โ”‚           38.93 / 39.19 ยฑ0.21 / 39.51 ms โ”‚    no change โ”‚
   โ”‚ QQuery 83 โ”‚           34.93 / 35.23 ยฑ0.27 / 35.70 ms โ”‚           34.86 / 35.25 ยฑ0.27 / 35.66 ms โ”‚    no change โ”‚
   โ”‚ QQuery 84 โ”‚           45.72 / 46.13 ยฑ0.26 / 46.47 ms โ”‚           45.90 / 47.24 ยฑ2.24 / 51.70 ms โ”‚    no change โ”‚
   โ”‚ QQuery 85 โ”‚        141.17 / 142.17 ยฑ0.86 / 143.69 ms โ”‚        140.18 / 140.77 ยฑ0.53 / 141.66 ms โ”‚    no change โ”‚
   โ”‚ QQuery 86 โ”‚           36.89 / 37.33 ยฑ0.50 / 38.25 ms โ”‚           37.11 / 38.36 ยฑ1.28 / 40.85 ms โ”‚    no change โ”‚
   โ”‚ QQuery 87 โ”‚              3.56 / 3.65 ยฑ0.12 / 3.87 ms โ”‚              3.49 / 3.60 ยฑ0.15 / 3.89 ms โ”‚    no change โ”‚
   โ”‚ QQuery 88 โ”‚         98.72 / 102.47 ยฑ6.71 / 115.86 ms โ”‚          99.15 / 99.88 ยฑ0.81 / 101.39 ms โ”‚    no change โ”‚
   โ”‚ QQuery 89 โ”‚        116.66 / 119.45 ยฑ4.52 / 128.47 ms โ”‚        116.19 / 119.90 ยฑ3.81 / 126.95 ms โ”‚    no change โ”‚
   โ”‚ QQuery 90 โ”‚           22.38 / 22.79 ยฑ0.21 / 22.98 ms โ”‚           22.30 / 22.99 ยฑ0.36 / 23.34 ms โ”‚    no change โ”‚
   โ”‚ QQuery 91 โ”‚           58.88 / 59.70 ยฑ0.82 / 60.93 ms โ”‚           58.18 / 59.22 ยฑ0.75 / 59.95 ms โ”‚    no change โ”‚
   โ”‚ QQuery 92 โ”‚           56.52 / 57.75 ยฑ1.72 / 61.17 ms โ”‚           55.57 / 56.20 ยฑ0.50 / 56.72 ms โ”‚    no change โ”‚
   โ”‚ QQuery 93 โ”‚        160.34 / 161.91 ยฑ1.18 / 163.67 ms โ”‚        160.74 / 162.01 ยฑ0.84 / 163.15 ms โ”‚    no change โ”‚
   โ”‚ QQuery 94 โ”‚           60.54 / 61.54 ยฑ1.02 / 63.39 ms โ”‚           60.36 / 60.91 ยฑ0.49 / 61.63 ms โ”‚    no change โ”‚
   โ”‚ QQuery 95 โ”‚        111.45 / 112.55 ยฑ0.70 / 113.35 ms โ”‚        110.57 / 112.56 ยฑ2.29 / 116.96 ms โ”‚    no change โ”‚
   โ”‚ QQuery 96 โ”‚           67.08 / 70.16 ยฑ2.97 / 75.61 ms โ”‚           68.97 / 69.80 ยฑ0.83 / 71.07 ms โ”‚    no change โ”‚
   โ”‚ QQuery 97 โ”‚        112.35 / 112.90 ยฑ0.47 / 113.72 ms โ”‚        112.76 / 114.28 ยฑ2.07 / 118.36 ms โ”‚    no change โ”‚
   โ”‚ QQuery 98 โ”‚        149.06 / 150.75 ยฑ0.97 / 151.77 ms โ”‚        149.26 / 151.48 ยฑ2.60 / 156.41 ms โ”‚    no change โ”‚
   โ”‚ QQuery 99 โ”‚ 10678.58 / 10724.50 ยฑ67.30 / 10857.74 ms โ”‚ 10683.14 / 10758.42 ยฑ50.28 / 10816.98 ms โ”‚    no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 29737.93ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 29730.96ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚   300.38ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚   300.31ms โ”‚
   โ”‚ Queries Faster                                    โ”‚          0 โ”‚
   โ”‚ Queries Slower                                    โ”‚          3 โ”‚
   โ”‚ Queries with No Change                            โ”‚         96 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **tpcds โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 150.0s |
   | Peak memory | 6.4 GiB |
   | Avg memory | 5.7 GiB |
   | CPU user | 235.8s |
   | CPU sys | 7.5s |
   | Peak spill | 0 B |
   
   **tpcds โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 150.0s |
   | Peak memory | 6.2 GiB |
   | Avg memory | 5.6 GiB |
   | CPU user | 235.4s |
   | CPU sys | 7.8s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364132649

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364085869)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark clickbench_extended.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                      HEAD โ”ƒ        speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 0  โ”‚        797.66 / 853.40 ยฑ39.29 / 900.75 ms โ”‚         808.65 / 820.69 ยฑ9.73 / 831.17 ms โ”‚     no change โ”‚
   โ”‚ QQuery 1  โ”‚         195.98 / 196.58 ยฑ0.42 / 197.09 ms โ”‚         191.18 / 196.75 ยฑ9.06 / 214.82 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚         483.96 / 486.37 ยฑ1.77 / 488.89 ms โ”‚         469.80 / 471.12 ยฑ1.49 / 473.96 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚         309.23 / 310.15 ยฑ0.78 / 311.28 ms โ”‚         310.35 / 311.97 ยฑ1.17 / 313.85 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚        644.80 / 660.17 ยฑ12.74 / 674.51 ms โ”‚         676.06 / 679.69 ยฑ3.13 / 684.24 ms โ”‚     no change โ”‚
   โ”‚ QQuery 5  โ”‚ 10051.91 / 10736.25 ยฑ419.67 / 11088.87 ms โ”‚ 10322.42 / 10587.51 ยฑ318.51 / 11180.06 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚     1007.70 / 1037.19 ยฑ40.49 / 1116.14 ms โ”‚     1015.31 / 1051.43 ยฑ20.04 / 1073.14 ms โ”‚     no change โ”‚
   โ”‚ QQuery 7  โ”‚        724.57 / 747.81 ยฑ17.94 / 765.80 ms โ”‚        789.05 / 810.50 ยฑ17.73 / 833.33 ms โ”‚  1.08x slower โ”‚
   โ”‚ QQuery 8  โ”‚        373.82 / 392.17 ยฑ16.91 / 412.81 ms โ”‚        380.66 / 407.28 ยฑ45.96 / 498.95 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚     2842.48 / 2894.02 ยฑ33.39 / 2938.15 ms โ”‚     2894.61 / 2966.84 ยฑ44.95 / 3015.09 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚        642.76 / 662.65 ยฑ19.93 / 700.04 ms โ”‚         644.55 / 650.56 ยฑ5.41 / 657.41 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚     2138.19 / 2202.58 ยฑ43.68 / 2252.32 ms โ”‚     2259.43 / 2295.05 ยฑ18.65 / 2313.83 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚        193.24 / 226.41 ยฑ55.03 / 334.97 ms โ”‚        198.97 / 214.17 ยฑ15.18 / 240.03 ms โ”‚ +1.06x faster โ”‚
   โ”‚ QQuery 13 โ”‚         551.83 / 556.84 ยฑ4.05 / 562.67 ms โ”‚        556.01 / 580.16 ยฑ20.66 / 617.95 ms โ”‚     no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 21962.59ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 22043.71ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚  1568.76ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚  1574.55ms โ”‚
   โ”‚ Queries Faster                                    โ”‚          1 โ”‚
   โ”‚ Queries Slower                                    โ”‚          1 โ”‚
   โ”‚ Queries with No Change                            โ”‚         12 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **clickbench_extended โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 115.0s |
   | Peak memory | 29.6 GiB |
   | Avg memory | 23.9 GiB |
   | CPU user | 1056.1s |
   | CPU sys | 43.2s |
   | Peak spill | 0 B |
   
   **clickbench_extended โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 115.0s |
   | Peak memory | 31.2 GiB |
   | Avg memory | 23.6 GiB |
   | CPU user | 1055.9s |
   | CPU sys | 43.4s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364606606

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364579262)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark tpch_sf1.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                           HEAD โ”ƒ speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 1  โ”‚ 40.19 / 40.85 ยฑ1.00 / 42.85 ms โ”‚     39.69 / 40.49 ยฑ1.12 / 42.69 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚ 21.01 / 21.34 ยฑ0.26 / 21.80 ms โ”‚     20.72 / 21.25 ยฑ0.69 / 22.59 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚ 36.76 / 38.90 ยฑ1.56 / 41.13 ms โ”‚     38.45 / 38.79 ยฑ0.50 / 39.79 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚ 18.34 / 18.51 ยฑ0.11 / 18.62 ms โ”‚     18.38 / 18.52 ยฑ0.09 / 18.65 ms โ”‚     no change โ”‚
   โ”‚ QQuery 5  โ”‚ 43.92 / 45.90 ยฑ1.44 / 47.69 ms โ”‚     44.27 / 45.78 ยฑ0.76 / 46.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚ 17.24 / 17.71 ยฑ0.53 / 18.75 ms โ”‚     17.02 / 17.67 ยฑ0.68 / 18.57 ms โ”‚     no change โ”‚
   โ”‚ QQuery 7  โ”‚ 50.30 / 51.95 ยฑ1.12 / 53.63 ms โ”‚     50.10 / 52.08 ยฑ1.29 / 54.10 ms โ”‚     no change โ”‚
   โ”‚ QQuery 8  โ”‚ 47.02 / 48.25 ยฑ1.15 / 50.10 ms โ”‚     47.12 / 47.97 ยฑ1.40 / 50.75 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚ 51.89 / 52.24 ยฑ0.18 / 52.43 ms โ”‚     51.90 / 53.24 ยฑ1.03 / 54.36 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚ 65.72 / 66.15 ยฑ0.37 / 66.58 ms โ”‚     65.60 / 66.85 ยฑ1.29 / 69.19 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚ 14.08 / 14.25 ยฑ0.16 / 14.52 ms โ”‚     14.30 / 14.37 ยฑ0.09 / 14.55 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚ 26.51 / 27.28 ยฑ1.03 / 29.30 ms โ”‚     26.22 / 26.70 ยฑ0.66 / 27.99 ms โ”‚     no change โ”‚
   โ”‚ QQuery 13 โ”‚ 34.53 / 35.80 ยฑ0.94 / 36.90 ms โ”‚     35.91 / 36.41 ยฑ0.47 / 37.28 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚ 26.71 / 27.03 ยฑ0.19 / 27.25 ms โ”‚     26.55 / 26.88 ยฑ0.20 / 27.12 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚ 33.18 / 33.67 ยฑ0.67 / 34.98 ms โ”‚     33.05 / 33.55 ยฑ0.68 / 34.88 ms โ”‚     no change โ”‚
   โ”‚ QQuery 16 โ”‚ 15.54 / 15.64 ยฑ0.07 / 15.74 ms โ”‚     15.26 / 15.41 ยฑ0.08 / 15.52 ms โ”‚     no change โ”‚
   โ”‚ QQuery 17 โ”‚ 76.51 / 77.34 ยฑ0.72 / 78.49 ms โ”‚     64.50 / 64.91 ยฑ0.30 / 65.41 ms โ”‚ +1.19x faster โ”‚
   โ”‚ QQuery 18 โ”‚ 69.26 / 69.58 ยฑ0.27 / 70.02 ms โ”‚     67.49 / 68.84 ยฑ0.84 / 69.87 ms โ”‚     no change โ”‚
   โ”‚ QQuery 19 โ”‚ 37.71 / 38.51 ยฑ1.13 / 40.64 ms โ”‚     37.32 / 37.63 ยฑ0.23 / 37.97 ms โ”‚     no change โ”‚
   โ”‚ QQuery 20 โ”‚ 38.89 / 39.22 ยฑ0.20 / 39.43 ms โ”‚     39.06 / 39.12 ยฑ0.05 / 39.18 ms โ”‚     no change โ”‚
   โ”‚ QQuery 21 โ”‚ 60.52 / 62.38 ยฑ0.97 / 63.13 ms โ”‚     60.34 / 61.99 ยฑ1.35 / 64.35 ms โ”‚     no change โ”‚
   โ”‚ QQuery 22 โ”‚ 23.65 / 23.92 ยฑ0.22 / 24.19 ms โ”‚     23.99 / 24.22 ยฑ0.22 / 24.55 ms โ”‚     no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ          โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 866.43ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 852.67ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚  39.38ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚  38.76ms โ”‚
   โ”‚ Queries Faster                                    โ”‚        1 โ”‚
   โ”‚ Queries Slower                                    โ”‚        0 โ”‚
   โ”‚ Queries with No Change                            โ”‚       21 โ”‚
   โ”‚ Queries with Failure                              โ”‚        0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **tpch โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 5.0s |
   | Peak memory | 5.4 GiB |
   | Avg memory | 5.0 GiB |
   | CPU user | 32.3s |
   | CPU sys | 2.2s |
   | Peak spill | 0 B |
   
   **tpch โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 5.0s |
   | Peak memory | 5.6 GiB |
   | Avg memory | 5.0 GiB |
   | CPU user | 31.4s |
   | CPU sys | 2.3s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "comphead (via GitHub)" <gi...@apache.org>.
comphead commented on code in PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#discussion_r3176139509


##########
datafusion/physical-plan/src/aggregates/group_values/single_group_by/primitive.rs:
##########
@@ -182,21 +174,30 @@ where
 
         let array: PrimitiveArray<T> = match emit_to {
             EmitTo::All => {
-                self.map.clear();
-                build_primitive(std::mem::take(&mut self.values), self.null_group.take())
+                let mut values = vec![T::default_value(); self.num_groups];

Review Comment:
   can `values` be reused? ๐Ÿค” 



-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363663561

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4363624206)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark clickbench_extended.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                     HEAD โ”ƒ        speed-up-single-primitive-group-by โ”ƒ    Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 0  โ”‚        761.11 / 776.76 ยฑ9.10 / 785.35 ms โ”‚         759.45 / 768.43 ยฑ4.62 / 771.94 ms โ”‚ no change โ”‚
   โ”‚ QQuery 1  โ”‚        195.96 / 196.86 ยฑ0.90 / 198.38 ms โ”‚         192.28 / 195.33 ยฑ4.81 / 204.89 ms โ”‚ no change โ”‚
   โ”‚ QQuery 2  โ”‚        481.97 / 485.79 ยฑ2.40 / 489.53 ms โ”‚         464.88 / 470.65 ยฑ4.64 / 478.86 ms โ”‚ no change โ”‚
   โ”‚ QQuery 3  โ”‚        309.57 / 310.76 ยฑ1.02 / 312.54 ms โ”‚         308.28 / 309.42 ยฑ1.39 / 312.14 ms โ”‚ no change โ”‚
   โ”‚ QQuery 4  โ”‚        658.83 / 666.69 ยฑ6.29 / 674.05 ms โ”‚         638.36 / 653.19 ยฑ9.01 / 661.71 ms โ”‚ no change โ”‚
   โ”‚ QQuery 5  โ”‚ 9892.52 / 10464.78 ยฑ405.91 / 10928.02 ms โ”‚ 10409.31 / 10612.58 ยฑ277.65 / 11161.46 ms โ”‚ no change โ”‚
   โ”‚ QQuery 6  โ”‚    1016.03 / 1028.79 ยฑ15.83 / 1059.55 ms โ”‚     1005.27 / 1014.08 ยฑ10.14 / 1033.49 ms โ”‚ no change โ”‚
   โ”‚ QQuery 7  โ”‚       718.21 / 756.12 ยฑ26.17 / 789.75 ms โ”‚        718.27 / 734.72 ยฑ11.99 / 749.32 ms โ”‚ no change โ”‚
   โ”‚ QQuery 8  โ”‚        373.20 / 376.62 ยฑ3.87 / 383.06 ms โ”‚         369.69 / 380.92 ยฑ8.91 / 395.83 ms โ”‚ no change โ”‚
   โ”‚ QQuery 9  โ”‚    2903.39 / 2928.79 ยฑ18.68 / 2956.49 ms โ”‚     2849.58 / 2893.42 ยฑ36.82 / 2946.36 ms โ”‚ no change โ”‚
   โ”‚ QQuery 10 โ”‚        645.52 / 656.89 ยฑ7.92 / 665.57 ms โ”‚        628.77 / 647.04 ยฑ16.96 / 677.08 ms โ”‚ no change โ”‚
   โ”‚ QQuery 11 โ”‚    2195.70 / 2208.50 ยฑ11.75 / 2227.74 ms โ”‚      2179.94 / 2191.70 ยฑ6.21 / 2197.36 ms โ”‚ no change โ”‚
   โ”‚ QQuery 12 โ”‚       195.44 / 214.22 ยฑ32.29 / 278.65 ms โ”‚        181.80 / 212.22 ยฑ57.41 / 326.99 ms โ”‚ no change โ”‚
   โ”‚ QQuery 13 โ”‚       559.63 / 571.42 ยฑ11.80 / 586.57 ms โ”‚         554.67 / 558.58 ยฑ4.02 / 563.73 ms โ”‚ no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 21642.99ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 21642.27ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚  1545.93ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚  1545.88ms โ”‚
   โ”‚ Queries Faster                                    โ”‚          0 โ”‚
   โ”‚ Queries Slower                                    โ”‚          0 โ”‚
   โ”‚ Queries with No Change                            โ”‚         14 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **clickbench_extended โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 110.0s |
   | Peak memory | 31.1 GiB |
   | Avg memory | 24.1 GiB |
   | CPU user | 1045.8s |
   | CPU sys | 43.2s |
   | Peak spill | 0 B |
   
   **clickbench_extended โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 110.0s |
   | Peak memory | 31.9 GiB |
   | Avg memory | 24.0 GiB |
   | CPU user | 1042.9s |
   | CPU sys | 42.2s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363628459

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4363624010)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4363624010-1972-t829t 6.12.55+ #1 SMP Sun Feb  1 08:59:41 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (58beeff7c2416a3e4da2644b320ff57f4c3e4016) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..58beeff7c2416a3e4da2644b320ff57f4c3e4016) using: tpch
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4360976823

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4360963663)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4360963663-1965-7r9sb 6.12.55+ #1 SMP Sun Feb  1 08:59:41 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (b79d7872b5b6070200a1b0838bf434ed6b3a6b87) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..b79d7872b5b6070200a1b0838bf434ed6b3a6b87) using: tpcds
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364085869

   run benchmark clickbench_extended


-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364615431

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364579262)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark clickbench_partitioned.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                  HEAD โ”ƒ    speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 0  โ”‚          1.22 / 4.62 ยฑ6.72 / 18.06 ms โ”‚          1.20 / 4.53 ยฑ6.61 / 17.74 ms โ”‚     no change โ”‚
   โ”‚ QQuery 1  โ”‚        12.52 / 12.61 ยฑ0.07 / 12.71 ms โ”‚        12.49 / 12.94 ยฑ0.26 / 13.31 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚        37.47 / 38.01 ยฑ0.32 / 38.35 ms โ”‚        36.40 / 36.63 ยฑ0.20 / 36.93 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚        31.36 / 31.91 ยฑ0.79 / 33.46 ms โ”‚        30.89 / 31.05 ยฑ0.16 / 31.31 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚     230.32 / 233.52 ยฑ2.89 / 238.53 ms โ”‚     227.27 / 228.48 ยฑ1.03 / 230.29 ms โ”‚     no change โ”‚
   โ”‚ QQuery 5  โ”‚     279.21 / 280.89 ยฑ1.02 / 282.25 ms โ”‚     277.62 / 278.28 ยฑ0.90 / 280.04 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚           6.85 / 7.41 ยฑ0.45 / 7.95 ms โ”‚           6.42 / 7.00 ยฑ0.63 / 7.87 ms โ”‚ +1.06x faster โ”‚
   โ”‚ QQuery 7  โ”‚        13.52 / 13.61 ยฑ0.09 / 13.78 ms โ”‚        14.07 / 14.19 ยฑ0.15 / 14.47 ms โ”‚     no change โ”‚
   โ”‚ QQuery 8  โ”‚     314.41 / 317.29 ยฑ1.67 / 318.97 ms โ”‚     311.57 / 313.54 ยฑ1.18 / 315.29 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚     448.24 / 452.76 ยฑ4.16 / 458.87 ms โ”‚     433.55 / 439.87 ยฑ3.64 / 443.79 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚        72.46 / 74.18 ยฑ1.44 / 76.80 ms โ”‚        72.24 / 73.26 ยฑ0.90 / 74.39 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚        84.95 / 85.55 ยฑ0.56 / 86.47 ms โ”‚        82.97 / 83.65 ยฑ0.44 / 84.23 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚     272.78 / 275.60 ยฑ2.70 / 280.58 ms โ”‚     270.30 / 274.10 ยฑ4.79 / 283.47 ms โ”‚     no change โ”‚
   โ”‚ QQuery 13 โ”‚     384.68 / 389.27 ยฑ4.94 / 396.16 ms โ”‚     379.52 / 390.13 ยฑ6.71 / 398.45 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚     281.28 / 283.97 ยฑ2.76 / 288.17 ms โ”‚     281.27 / 284.36 ยฑ2.58 / 288.65 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚     271.12 / 275.95 ยฑ3.01 / 279.75 ms โ”‚     275.90 / 281.57 ยฑ6.28 / 292.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 16 โ”‚     604.79 / 610.06 ยฑ3.26 / 613.73 ms โ”‚     607.34 / 614.00 ยฑ6.75 / 626.36 ms โ”‚     no change โ”‚
   โ”‚ QQuery 17 โ”‚     611.02 / 614.95 ยฑ3.91 / 622.21 ms โ”‚     606.87 / 612.91 ยฑ4.19 / 619.12 ms โ”‚     no change โ”‚
   โ”‚ QQuery 18 โ”‚  1229.59 / 1237.07 ยฑ8.93 / 1254.63 ms โ”‚  1226.93 / 1232.75 ยฑ6.09 / 1243.35 ms โ”‚     no change โ”‚
   โ”‚ QQuery 19 โ”‚        28.37 / 30.19 ยฑ2.87 / 35.89 ms โ”‚       28.40 / 36.13 ยฑ10.15 / 55.25 ms โ”‚  1.20x slower โ”‚
   โ”‚ QQuery 20 โ”‚     515.39 / 528.20 ยฑ8.62 / 539.61 ms โ”‚     519.37 / 523.04 ยฑ3.87 / 530.49 ms โ”‚     no change โ”‚
   โ”‚ QQuery 21 โ”‚     592.74 / 600.75 ยฑ6.42 / 611.54 ms โ”‚     595.37 / 598.19 ยฑ2.24 / 601.13 ms โ”‚     no change โ”‚
   โ”‚ QQuery 22 โ”‚  1065.60 / 1068.92 ยฑ1.75 / 1070.47 ms โ”‚  1063.99 / 1072.08 ยฑ5.51 / 1080.15 ms โ”‚     no change โ”‚
   โ”‚ QQuery 23 โ”‚ 3327.90 / 3352.34 ยฑ22.53 / 3389.69 ms โ”‚ 3283.12 / 3322.68 ยฑ33.38 / 3375.30 ms โ”‚     no change โ”‚
   โ”‚ QQuery 24 โ”‚        42.28 / 47.88 ยฑ9.99 / 67.83 ms โ”‚        41.71 / 42.22 ยฑ0.42 / 42.93 ms โ”‚ +1.13x faster โ”‚
   โ”‚ QQuery 25 โ”‚     113.19 / 114.16 ยฑ0.95 / 115.97 ms โ”‚     113.87 / 118.90 ยฑ7.38 / 133.38 ms โ”‚     no change โ”‚
   โ”‚ QQuery 26 โ”‚        42.34 / 43.77 ยฑ2.66 / 49.09 ms โ”‚        42.41 / 43.11 ยฑ0.59 / 44.04 ms โ”‚     no change โ”‚
   โ”‚ QQuery 27 โ”‚     668.29 / 671.23 ยฑ1.65 / 673.00 ms โ”‚    663.71 / 676.14 ยฑ11.58 / 695.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 28 โ”‚ 2995.11 / 3017.19 ยฑ16.73 / 3042.13 ms โ”‚ 2998.27 / 3051.33 ยฑ47.73 / 3134.51 ms โ”‚     no change โ”‚
   โ”‚ QQuery 29 โ”‚        42.84 / 44.87 ยฑ3.59 / 52.04 ms โ”‚        42.13 / 47.82 ยฑ5.46 / 55.62 ms โ”‚  1.07x slower โ”‚
   โ”‚ QQuery 30 โ”‚     302.08 / 308.87 ยฑ5.90 / 318.51 ms โ”‚     302.45 / 308.03 ยฑ4.89 / 316.49 ms โ”‚     no change โ”‚
   โ”‚ QQuery 31 โ”‚     293.53 / 301.84 ยฑ6.34 / 311.82 ms โ”‚     297.43 / 304.67 ยฑ6.76 / 314.22 ms โ”‚     no change โ”‚
   โ”‚ QQuery 32 โ”‚     964.13 / 972.52 ยฑ8.06 / 983.14 ms โ”‚   964.33 / 982.12 ยฑ17.39 / 1006.80 ms โ”‚     no change โ”‚
   โ”‚ QQuery 33 โ”‚  1405.05 / 1413.79 ยฑ6.29 / 1422.31 ms โ”‚ 1406.86 / 1428.01 ยฑ16.64 / 1455.78 ms โ”‚     no change โ”‚
   โ”‚ QQuery 34 โ”‚  1422.78 / 1437.60 ยฑ9.30 / 1451.35 ms โ”‚ 1407.26 / 1433.42 ยฑ14.65 / 1448.10 ms โ”‚     no change โ”‚
   โ”‚ QQuery 35 โ”‚     278.53 / 284.10 ยฑ5.91 / 294.30 ms โ”‚    259.73 / 283.82 ยฑ23.09 / 323.11 ms โ”‚     no change โ”‚
   โ”‚ QQuery 36 โ”‚        61.01 / 66.91 ยฑ3.48 / 70.70 ms โ”‚        63.26 / 70.35 ยฑ4.37 / 75.75 ms โ”‚  1.05x slower โ”‚
   โ”‚ QQuery 37 โ”‚        35.00 / 38.26 ยฑ4.59 / 47.36 ms โ”‚        35.53 / 38.62 ยฑ5.62 / 49.85 ms โ”‚     no change โ”‚
   โ”‚ QQuery 38 โ”‚        40.63 / 42.64 ยฑ1.37 / 44.24 ms โ”‚        40.11 / 42.47 ยฑ1.39 / 44.00 ms โ”‚     no change โ”‚
   โ”‚ QQuery 39 โ”‚     125.33 / 134.12 ยฑ7.08 / 144.53 ms โ”‚     129.50 / 135.42 ยฑ4.67 / 141.98 ms โ”‚     no change โ”‚
   โ”‚ QQuery 40 โ”‚        14.12 / 14.39 ยฑ0.27 / 14.86 ms โ”‚        14.25 / 14.48 ยฑ0.20 / 14.84 ms โ”‚     no change โ”‚
   โ”‚ QQuery 41 โ”‚        13.56 / 14.48 ยฑ1.63 / 17.73 ms โ”‚        13.93 / 14.09 ยฑ0.11 / 14.21 ms โ”‚     no change โ”‚
   โ”‚ QQuery 42 โ”‚        13.30 / 13.45 ยฑ0.12 / 13.59 ms โ”‚        13.94 / 14.03 ยฑ0.06 / 14.12 ms โ”‚     no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 19801.69ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 19814.41ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚   460.50ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚   460.80ms โ”‚
   โ”‚ Queries Faster                                    โ”‚          2 โ”‚
   โ”‚ Queries Slower                                    โ”‚          3 โ”‚
   โ”‚ Queries with No Change                            โ”‚         38 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **clickbench_partitioned โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 100.0s |
   | Peak memory | 30.1 GiB |
   | Avg memory | 23.1 GiB |
   | CPU user | 1049.3s |
   | CPU sys | 60.8s |
   | Peak spill | 0 B |
   
   **clickbench_partitioned โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 100.0s |
   | Peak memory | 29.9 GiB |
   | Avg memory | 23.1 GiB |
   | CPU user | 1046.0s |
   | CPU sys | 62.4s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan closed pull request #21977: perf: speed up primitive group value interning, use less memory
URL: https://github.com/apache/datafusion/pull/21977


-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363628513

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4363624010)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4363624010-1970-jm72p 6.12.55+ #1 SMP Sun Feb  1 08:59:41 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (58beeff7c2416a3e4da2644b320ff57f4c3e4016) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..58beeff7c2416a3e4da2644b320ff57f4c3e4016) using: clickbench_partitioned
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364381019

   Ill test the dense path some more


-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364615275

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364579262)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark tpcds_sf1.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                     HEAD โ”ƒ       speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 1  โ”‚              7.40 / 8.03 ยฑ0.81 / 9.64 ms โ”‚              7.67 / 8.10 ยฑ0.75 / 9.59 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚        146.00 / 146.32 ยฑ0.24 / 146.74 ms โ”‚        141.50 / 142.02 ยฑ0.40 / 142.50 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚        113.24 / 115.92 ยฑ2.16 / 119.77 ms โ”‚        113.63 / 114.57 ยฑ0.52 / 115.05 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚    1284.67 / 1303.94 ยฑ15.00 / 1326.63 ms โ”‚    1304.59 / 1375.49 ยฑ54.36 / 1444.21 ms โ”‚  1.05x slower โ”‚
   โ”‚ QQuery 5  โ”‚        173.96 / 176.21 ยฑ1.65 / 178.55 ms โ”‚        176.53 / 179.11 ยฑ1.64 / 181.55 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚       129.04 / 141.00 ยฑ14.29 / 168.51 ms โ”‚       132.82 / 143.40 ยฑ14.44 / 171.43 ms โ”‚     no change โ”‚
   โ”‚ QQuery 7  โ”‚        320.22 / 322.56 ยฑ2.99 / 328.46 ms โ”‚        316.80 / 321.16 ยฑ3.62 / 326.68 ms โ”‚     no change โ”‚
   โ”‚ QQuery 8  โ”‚        112.60 / 114.37 ยฑ1.48 / 116.78 ms โ”‚        112.59 / 113.14 ยฑ0.57 / 114.17 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚         97.51 / 105.19 ยฑ9.28 / 123.47 ms โ”‚          94.89 / 98.00 ยฑ3.01 / 102.27 ms โ”‚ +1.07x faster โ”‚
   โ”‚ QQuery 10 โ”‚        103.28 / 104.41 ยฑ1.41 / 107.08 ms โ”‚        101.92 / 102.67 ยฑ0.54 / 103.31 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚        888.37 / 898.43 ยฑ9.37 / 914.37 ms โ”‚      881.25 / 949.72 ยฑ52.27 / 1014.35 ms โ”‚  1.06x slower โ”‚
   โ”‚ QQuery 12 โ”‚           45.25 / 45.90 ยฑ0.36 / 46.27 ms โ”‚           45.73 / 46.19 ยฑ0.40 / 46.75 ms โ”‚     no change โ”‚
   โ”‚ QQuery 13 โ”‚        390.22 / 402.11 ยฑ6.20 / 408.44 ms โ”‚        386.96 / 393.49 ยฑ6.83 / 406.17 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚    1033.06 / 1046.20 ยฑ10.44 / 1061.18 ms โ”‚     1042.25 / 1048.78 ยฑ6.56 / 1058.22 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚           14.83 / 15.43 ยฑ0.39 / 16.02 ms โ”‚           15.20 / 15.89 ยฑ0.45 / 16.60 ms โ”‚     no change โ”‚
   โ”‚ QQuery 16 โ”‚              7.37 / 7.50 ยฑ0.21 / 7.92 ms โ”‚              8.43 / 8.56 ยฑ0.14 / 8.82 ms โ”‚  1.14x slower โ”‚
   โ”‚ QQuery 17 โ”‚        204.57 / 208.42 ยฑ2.40 / 210.65 ms โ”‚        202.14 / 204.37 ยฑ2.22 / 208.11 ms โ”‚     no change โ”‚
   โ”‚ QQuery 18 โ”‚        128.23 / 129.40 ยฑ1.35 / 131.95 ms โ”‚        122.11 / 124.51 ยฑ1.54 / 126.98 ms โ”‚     no change โ”‚
   โ”‚ QQuery 19 โ”‚        158.44 / 161.85 ยฑ1.74 / 163.12 ms โ”‚        155.89 / 160.12 ยฑ4.03 / 166.18 ms โ”‚     no change โ”‚
   โ”‚ QQuery 20 โ”‚           14.06 / 14.35 ยฑ0.17 / 14.59 ms โ”‚           12.70 / 13.17 ยฑ0.44 / 13.90 ms โ”‚ +1.09x faster โ”‚
   โ”‚ QQuery 21 โ”‚           19.81 / 20.22 ยฑ0.30 / 20.57 ms โ”‚           18.74 / 18.82 ยฑ0.06 / 18.90 ms โ”‚ +1.07x faster โ”‚
   โ”‚ QQuery 22 โ”‚        488.54 / 493.15 ยฑ3.89 / 499.05 ms โ”‚        473.72 / 481.51 ยฑ7.12 / 492.77 ms โ”‚     no change โ”‚
   โ”‚ QQuery 23 โ”‚    1016.94 / 1056.94 ยฑ35.62 / 1100.68 ms โ”‚    1017.84 / 1030.23 ยฑ10.66 / 1049.34 ms โ”‚     no change โ”‚
   โ”‚ QQuery 24 โ”‚        670.19 / 672.89 ยฑ2.86 / 677.58 ms โ”‚        638.53 / 643.08 ยฑ4.13 / 650.89 ms โ”‚     no change โ”‚
   โ”‚ QQuery 25 โ”‚        307.71 / 317.87 ยฑ6.69 / 325.30 ms โ”‚        302.90 / 306.77 ยฑ2.52 / 309.99 ms โ”‚     no change โ”‚
   โ”‚ QQuery 26 โ”‚           77.20 / 77.53 ยฑ0.36 / 78.18 ms โ”‚           75.52 / 76.07 ยฑ0.57 / 77.15 ms โ”‚     no change โ”‚
   โ”‚ QQuery 27 โ”‚              7.18 / 7.32 ยฑ0.17 / 7.65 ms โ”‚              8.09 / 8.15 ยฑ0.07 / 8.28 ms โ”‚  1.11x slower โ”‚
   โ”‚ QQuery 28 โ”‚        148.78 / 153.50 ยฑ5.17 / 163.02 ms โ”‚        146.73 / 149.17 ยฑ1.91 / 152.19 ms โ”‚     no change โ”‚
   โ”‚ QQuery 29 โ”‚        244.90 / 251.69 ยฑ7.54 / 266.41 ms โ”‚        245.40 / 250.27 ยฑ3.24 / 255.23 ms โ”‚     no change โ”‚
   โ”‚ QQuery 30 โ”‚           41.82 / 42.26 ยฑ0.32 / 42.65 ms โ”‚           42.75 / 43.41 ยฑ0.48 / 44.19 ms โ”‚     no change โ”‚
   โ”‚ QQuery 31 โ”‚        165.33 / 169.31 ยฑ5.18 / 179.47 ms โ”‚        167.87 / 169.90 ยฑ1.65 / 172.76 ms โ”‚     no change โ”‚
   โ”‚ QQuery 32 โ”‚           13.26 / 13.48 ยฑ0.21 / 13.81 ms โ”‚           13.73 / 14.09 ยฑ0.33 / 14.63 ms โ”‚     no change โ”‚
   โ”‚ QQuery 33 โ”‚        138.35 / 141.60 ยฑ3.79 / 148.95 ms โ”‚        139.20 / 141.28 ยฑ1.82 / 143.92 ms โ”‚     no change โ”‚
   โ”‚ QQuery 34 โ”‚              6.90 / 7.06 ยฑ0.20 / 7.46 ms โ”‚              7.07 / 7.22 ยฑ0.18 / 7.57 ms โ”‚     no change โ”‚
   โ”‚ QQuery 35 โ”‚        100.32 / 100.87 ยฑ0.38 / 101.50 ms โ”‚          99.09 / 99.92 ยฑ0.64 / 100.87 ms โ”‚     no change โ”‚
   โ”‚ QQuery 36 โ”‚              6.78 / 7.00 ยฑ0.14 / 7.16 ms โ”‚              6.94 / 7.13 ยฑ0.15 / 7.35 ms โ”‚     no change โ”‚
   โ”‚ QQuery 37 โ”‚              8.93 / 9.03 ยฑ0.07 / 9.09 ms โ”‚              8.16 / 8.34 ยฑ0.13 / 8.52 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 38 โ”‚          85.89 / 90.40 ยฑ5.32 / 100.75 ms โ”‚           88.49 / 93.50 ยฑ3.15 / 96.42 ms โ”‚     no change โ”‚
   โ”‚ QQuery 39 โ”‚        117.42 / 119.93 ยฑ4.04 / 127.95 ms โ”‚        115.67 / 118.92 ยฑ4.60 / 128.04 ms โ”‚     no change โ”‚
   โ”‚ QQuery 40 โ”‚          93.95 / 96.56 ยฑ3.78 / 104.03 ms โ”‚         96.49 / 100.95 ยฑ2.81 / 105.37 ms โ”‚     no change โ”‚
   โ”‚ QQuery 41 โ”‚           14.53 / 14.79 ยฑ0.25 / 15.25 ms โ”‚           14.34 / 14.56 ยฑ0.27 / 15.06 ms โ”‚     no change โ”‚
   โ”‚ QQuery 42 โ”‚        107.37 / 109.18 ยฑ2.60 / 114.32 ms โ”‚        106.36 / 107.41 ยฑ1.01 / 109.15 ms โ”‚     no change โ”‚
   โ”‚ QQuery 43 โ”‚              5.88 / 6.04 ยฑ0.20 / 6.43 ms โ”‚              6.15 / 6.44 ยฑ0.17 / 6.63 ms โ”‚  1.07x slower โ”‚
   โ”‚ QQuery 44 โ”‚           11.47 / 11.95 ยฑ0.38 / 12.39 ms โ”‚           11.90 / 12.22 ยฑ0.22 / 12.49 ms โ”‚     no change โ”‚
   โ”‚ QQuery 45 โ”‚           46.46 / 46.80 ยฑ0.21 / 47.02 ms โ”‚           46.32 / 47.17 ยฑ0.90 / 48.85 ms โ”‚     no change โ”‚
   โ”‚ QQuery 46 โ”‚              9.09 / 9.24 ยฑ0.14 / 9.48 ms โ”‚              9.44 / 9.50 ยฑ0.08 / 9.65 ms โ”‚     no change โ”‚
   โ”‚ QQuery 47 โ”‚       681.87 / 723.34 ยฑ41.56 / 790.12 ms โ”‚       692.21 / 718.36 ยฑ15.53 / 733.52 ms โ”‚     no change โ”‚
   โ”‚ QQuery 48 โ”‚        279.60 / 289.54 ยฑ8.41 / 303.01 ms โ”‚        272.97 / 281.75 ยฑ4.80 / 287.52 ms โ”‚     no change โ”‚
   โ”‚ QQuery 49 โ”‚        244.79 / 247.87 ยฑ2.02 / 250.98 ms โ”‚        248.18 / 249.86 ยฑ1.42 / 252.44 ms โ”‚     no change โ”‚
   โ”‚ QQuery 50 โ”‚        182.73 / 186.46 ยฑ3.00 / 190.23 ms โ”‚        198.33 / 202.82 ยฑ4.53 / 210.56 ms โ”‚  1.09x slower โ”‚
   โ”‚ QQuery 51 โ”‚        171.53 / 173.47 ยฑ1.40 / 175.90 ms โ”‚        173.93 / 175.69 ยฑ1.16 / 177.06 ms โ”‚     no change โ”‚
   โ”‚ QQuery 52 โ”‚        106.31 / 107.66 ยฑ1.25 / 109.89 ms โ”‚        108.28 / 110.91 ยฑ3.39 / 117.61 ms โ”‚     no change โ”‚
   โ”‚ QQuery 53 โ”‚        102.56 / 105.36 ยฑ2.13 / 108.80 ms โ”‚        104.49 / 105.09 ยฑ0.56 / 106.10 ms โ”‚     no change โ”‚
   โ”‚ QQuery 54 โ”‚        150.63 / 153.99 ยฑ3.49 / 160.47 ms โ”‚        149.29 / 152.62 ยฑ2.92 / 157.84 ms โ”‚     no change โ”‚
   โ”‚ QQuery 55 โ”‚        106.84 / 108.80 ยฑ1.01 / 109.61 ms โ”‚        108.11 / 108.83 ยฑ0.86 / 110.39 ms โ”‚     no change โ”‚
   โ”‚ QQuery 56 โ”‚        139.85 / 141.02 ยฑ1.49 / 143.92 ms โ”‚        138.79 / 141.28 ยฑ2.32 / 145.04 ms โ”‚     no change โ”‚
   โ”‚ QQuery 57 โ”‚        164.73 / 167.80 ยฑ2.09 / 171.23 ms โ”‚        164.90 / 166.75 ยฑ2.44 / 171.53 ms โ”‚     no change โ”‚
   โ”‚ QQuery 58 โ”‚        243.44 / 246.35 ยฑ2.12 / 248.70 ms โ”‚        241.33 / 243.90 ยฑ1.49 / 245.55 ms โ”‚     no change โ”‚
   โ”‚ QQuery 59 โ”‚        195.50 / 197.00 ยฑ1.76 / 200.23 ms โ”‚        197.49 / 201.80 ยฑ2.92 / 206.46 ms โ”‚     no change โ”‚
   โ”‚ QQuery 60 โ”‚        140.93 / 142.51 ยฑ1.39 / 144.21 ms โ”‚        146.12 / 148.28 ยฑ2.77 / 153.68 ms โ”‚     no change โ”‚
   โ”‚ QQuery 61 โ”‚           14.19 / 14.37 ยฑ0.15 / 14.60 ms โ”‚           14.08 / 14.13 ยฑ0.05 / 14.21 ms โ”‚     no change โ”‚
   โ”‚ QQuery 62 โ”‚        857.13 / 866.67 ยฑ7.47 / 879.18 ms โ”‚       878.87 / 909.94 ยฑ23.83 / 941.07 ms โ”‚     no change โ”‚
   โ”‚ QQuery 63 โ”‚        101.94 / 103.37 ยฑ1.06 / 104.36 ms โ”‚        102.83 / 104.99 ยฑ2.56 / 109.74 ms โ”‚     no change โ”‚
   โ”‚ QQuery 64 โ”‚        627.59 / 634.15 ยฑ7.13 / 647.42 ms โ”‚        624.25 / 636.24 ยฑ6.37 / 642.90 ms โ”‚     no change โ”‚
   โ”‚ QQuery 65 โ”‚        241.63 / 251.12 ยฑ6.79 / 259.20 ms โ”‚       240.15 / 256.92 ยฑ10.75 / 269.55 ms โ”‚     no change โ”‚
   โ”‚ QQuery 66 โ”‚        214.94 / 223.97 ยฑ6.05 / 233.71 ms โ”‚       216.86 / 228.04 ยฑ12.16 / 251.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 67 โ”‚       286.96 / 295.34 ยฑ11.02 / 316.80 ms โ”‚       289.69 / 299.67 ยฑ11.52 / 321.52 ms โ”‚     no change โ”‚
   โ”‚ QQuery 68 โ”‚              8.61 / 8.85 ยฑ0.26 / 9.35 ms โ”‚              9.18 / 9.34 ยฑ0.26 / 9.87 ms โ”‚  1.06x slower โ”‚
   โ”‚ QQuery 69 โ”‚         99.13 / 101.92 ยฑ2.33 / 105.73 ms โ”‚          96.81 / 98.93 ยฑ2.44 / 103.38 ms โ”‚     no change โ”‚
   โ”‚ QQuery 70 โ”‚       309.30 / 321.67 ยฑ13.48 / 346.97 ms โ”‚       324.45 / 334.75 ยฑ11.40 / 356.77 ms โ”‚     no change โ”‚
   โ”‚ QQuery 71 โ”‚        136.98 / 138.01 ยฑ0.99 / 139.54 ms โ”‚        132.31 / 136.25 ยฑ3.83 / 142.91 ms โ”‚     no change โ”‚
   โ”‚ QQuery 72 โ”‚       535.42 / 556.49 ยฑ16.40 / 582.51 ms โ”‚        544.16 / 551.94 ยฑ8.39 / 568.24 ms โ”‚     no change โ”‚
   โ”‚ QQuery 73 โ”‚              7.28 / 7.45 ยฑ0.13 / 7.64 ms โ”‚              6.84 / 7.03 ยฑ0.23 / 7.48 ms โ”‚ +1.06x faster โ”‚
   โ”‚ QQuery 74 โ”‚       540.79 / 579.27 ยฑ37.90 / 638.05 ms โ”‚       556.21 / 610.49 ยฑ52.97 / 678.28 ms โ”‚  1.05x slower โ”‚
   โ”‚ QQuery 75 โ”‚        266.57 / 270.02 ยฑ2.38 / 273.35 ms โ”‚        272.15 / 275.95 ยฑ2.47 / 278.72 ms โ”‚     no change โ”‚
   โ”‚ QQuery 76 โ”‚        127.95 / 130.42 ยฑ1.98 / 133.35 ms โ”‚        132.15 / 134.99 ยฑ3.44 / 141.70 ms โ”‚     no change โ”‚
   โ”‚ QQuery 77 โ”‚        187.46 / 190.28 ยฑ1.50 / 191.78 ms โ”‚        191.04 / 192.39 ยฑ1.69 / 195.59 ms โ”‚     no change โ”‚
   โ”‚ QQuery 78 โ”‚        316.59 / 321.38 ยฑ2.79 / 324.18 ms โ”‚        312.61 / 320.89 ยฑ6.35 / 328.91 ms โ”‚     no change โ”‚
   โ”‚ QQuery 79 โ”‚        239.37 / 247.88 ยฑ6.26 / 253.71 ms โ”‚        232.05 / 238.30 ยฑ5.67 / 248.53 ms โ”‚     no change โ”‚
   โ”‚ QQuery 80 โ”‚        303.00 / 305.86 ยฑ1.98 / 308.44 ms โ”‚        300.66 / 303.89 ยฑ2.10 / 305.81 ms โ”‚     no change โ”‚
   โ”‚ QQuery 81 โ”‚           27.45 / 27.58 ยฑ0.10 / 27.73 ms โ”‚           25.68 / 26.04 ยฑ0.23 / 26.31 ms โ”‚ +1.06x faster โ”‚
   โ”‚ QQuery 82 โ”‚           40.39 / 40.88 ยฑ0.42 / 41.44 ms โ”‚           38.81 / 40.13 ยฑ0.83 / 40.87 ms โ”‚     no change โ”‚
   โ”‚ QQuery 83 โ”‚           36.73 / 38.05 ยฑ1.96 / 41.93 ms โ”‚           34.85 / 35.10 ยฑ0.21 / 35.47 ms โ”‚ +1.08x faster โ”‚
   โ”‚ QQuery 84 โ”‚           47.57 / 47.79 ยฑ0.18 / 48.07 ms โ”‚           45.88 / 46.06 ยฑ0.13 / 46.29 ms โ”‚     no change โ”‚
   โ”‚ QQuery 85 โ”‚        144.39 / 146.28 ยฑ3.07 / 152.39 ms โ”‚        139.20 / 141.05 ยฑ1.45 / 143.25 ms โ”‚     no change โ”‚
   โ”‚ QQuery 86 โ”‚           38.35 / 38.96 ยฑ0.51 / 39.78 ms โ”‚           36.99 / 38.01 ยฑ0.51 / 38.36 ms โ”‚     no change โ”‚
   โ”‚ QQuery 87 โ”‚              3.96 / 4.03 ยฑ0.08 / 4.18 ms โ”‚              3.51 / 3.59 ยฑ0.16 / 3.91 ms โ”‚ +1.12x faster โ”‚
   โ”‚ QQuery 88 โ”‚        101.92 / 102.18 ยฑ0.19 / 102.43 ms โ”‚          98.70 / 99.94 ยฑ0.82 / 101.05 ms โ”‚     no change โ”‚
   โ”‚ QQuery 89 โ”‚        119.88 / 123.82 ยฑ5.52 / 134.70 ms โ”‚        116.25 / 119.42 ยฑ5.19 / 129.68 ms โ”‚     no change โ”‚
   โ”‚ QQuery 90 โ”‚           23.48 / 23.85 ยฑ0.40 / 24.54 ms โ”‚           22.07 / 22.42 ยฑ0.32 / 22.85 ms โ”‚ +1.06x faster โ”‚
   โ”‚ QQuery 91 โ”‚           61.08 / 61.54 ยฑ0.34 / 62.02 ms โ”‚           58.53 / 59.31 ยฑ0.59 / 59.95 ms โ”‚     no change โ”‚
   โ”‚ QQuery 92 โ”‚           56.55 / 58.97 ยฑ2.63 / 63.93 ms โ”‚           55.75 / 57.51 ยฑ1.26 / 58.65 ms โ”‚     no change โ”‚
   โ”‚ QQuery 93 โ”‚        161.61 / 163.39 ยฑ1.03 / 164.56 ms โ”‚        161.02 / 163.62 ยฑ3.69 / 170.96 ms โ”‚     no change โ”‚
   โ”‚ QQuery 94 โ”‚           60.79 / 61.35 ยฑ0.36 / 61.78 ms โ”‚           60.58 / 61.58 ยฑ1.18 / 63.57 ms โ”‚     no change โ”‚
   โ”‚ QQuery 95 โ”‚        111.94 / 112.50 ยฑ0.39 / 112.89 ms โ”‚        110.09 / 111.52 ยฑ1.06 / 113.12 ms โ”‚     no change โ”‚
   โ”‚ QQuery 96 โ”‚           68.05 / 69.76 ยฑ1.86 / 73.26 ms โ”‚           67.97 / 70.98 ยฑ3.33 / 77.47 ms โ”‚     no change โ”‚
   โ”‚ QQuery 97 โ”‚        114.39 / 116.34 ยฑ1.28 / 118.09 ms โ”‚        112.30 / 113.53 ยฑ0.98 / 115.01 ms โ”‚     no change โ”‚
   โ”‚ QQuery 98 โ”‚        157.98 / 158.94 ยฑ1.01 / 160.37 ms โ”‚        150.05 / 152.78 ยฑ2.16 / 156.63 ms โ”‚     no change โ”‚
   โ”‚ QQuery 99 โ”‚ 10746.66 / 10892.86 ยฑ96.05 / 11001.78 ms โ”‚ 10793.65 / 10904.45 ยฑ84.03 / 10983.44 ms โ”‚     no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 30362.96ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 30468.60ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚   306.70ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚   307.76ms โ”‚
   โ”‚ Queries Faster                                    โ”‚          9 โ”‚
   โ”‚ Queries Slower                                    โ”‚          8 โ”‚
   โ”‚ Queries with No Change                            โ”‚         82 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **tpcds โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 155.0s |
   | Peak memory | 6.3 GiB |
   | Avg memory | 5.6 GiB |
   | CPU user | 241.4s |
   | CPU sys | 7.9s |
   | Peak spill | 0 B |
   
   **tpcds โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 155.0s |
   | Peak memory | 6.2 GiB |
   | Avg memory | 5.6 GiB |
   | CPU user | 241.0s |
   | CPU sys | 8.0s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363658816

   ๐Ÿค– Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4363624010)
   
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   
   <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 speed-up-single-primitive-group-by
   --------------------
   Benchmark clickbench_partitioned.json
   --------------------
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Query     โ”ƒ                                  HEAD โ”ƒ    speed-up-single-primitive-group-by โ”ƒ        Change โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ QQuery 0  โ”‚          1.21 / 4.68 ยฑ6.80 / 18.28 ms โ”‚          1.20 / 4.67 ยฑ6.80 / 18.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 1  โ”‚        12.67 / 13.13 ยฑ0.23 / 13.29 ms โ”‚        12.12 / 12.65 ยฑ0.28 / 12.90 ms โ”‚     no change โ”‚
   โ”‚ QQuery 2  โ”‚        36.95 / 37.21 ยฑ0.20 / 37.49 ms โ”‚        37.06 / 37.34 ยฑ0.21 / 37.63 ms โ”‚     no change โ”‚
   โ”‚ QQuery 3  โ”‚        32.09 / 33.03 ยฑ0.80 / 33.98 ms โ”‚        31.63 / 32.40 ยฑ0.95 / 34.02 ms โ”‚     no change โ”‚
   โ”‚ QQuery 4  โ”‚     233.91 / 237.57 ยฑ2.99 / 242.79 ms โ”‚     208.25 / 210.47 ยฑ1.72 / 212.96 ms โ”‚ +1.13x faster โ”‚
   โ”‚ QQuery 5  โ”‚     284.44 / 286.02 ยฑ1.76 / 289.37 ms โ”‚     281.01 / 283.36 ยฑ1.44 / 284.83 ms โ”‚     no change โ”‚
   โ”‚ QQuery 6  โ”‚           6.75 / 6.96 ยฑ0.16 / 7.20 ms โ”‚           6.53 / 7.02 ยฑ0.43 / 7.61 ms โ”‚     no change โ”‚
   โ”‚ QQuery 7  โ”‚        14.43 / 15.06 ยฑ0.96 / 16.97 ms โ”‚        13.60 / 13.70 ยฑ0.08 / 13.84 ms โ”‚ +1.10x faster โ”‚
   โ”‚ QQuery 8  โ”‚     325.78 / 328.25 ยฑ2.04 / 330.38 ms โ”‚     318.71 / 321.71 ยฑ2.20 / 324.79 ms โ”‚     no change โ”‚
   โ”‚ QQuery 9  โ”‚     453.98 / 465.61 ยฑ8.12 / 478.86 ms โ”‚     447.03 / 453.13 ยฑ4.74 / 460.26 ms โ”‚     no change โ”‚
   โ”‚ QQuery 10 โ”‚        74.96 / 75.90 ยฑ0.80 / 77.29 ms โ”‚        71.95 / 73.07 ยฑ0.79 / 74.39 ms โ”‚     no change โ”‚
   โ”‚ QQuery 11 โ”‚        85.43 / 86.43 ยฑ0.53 / 86.95 ms โ”‚        84.22 / 84.72 ยฑ0.54 / 85.69 ms โ”‚     no change โ”‚
   โ”‚ QQuery 12 โ”‚     276.27 / 280.26 ยฑ3.56 / 286.94 ms โ”‚     274.94 / 278.57 ยฑ4.14 / 284.83 ms โ”‚     no change โ”‚
   โ”‚ QQuery 13 โ”‚     389.36 / 398.67 ยฑ8.39 / 412.57 ms โ”‚     386.57 / 393.81 ยฑ4.99 / 401.21 ms โ”‚     no change โ”‚
   โ”‚ QQuery 14 โ”‚     289.09 / 291.76 ยฑ2.31 / 294.61 ms โ”‚     286.25 / 288.86 ยฑ2.09 / 292.24 ms โ”‚     no change โ”‚
   โ”‚ QQuery 15 โ”‚     280.43 / 283.54 ยฑ2.39 / 286.62 ms โ”‚     253.49 / 259.48 ยฑ7.97 / 274.37 ms โ”‚ +1.09x faster โ”‚
   โ”‚ QQuery 16 โ”‚     612.65 / 625.78 ยฑ8.39 / 637.05 ms โ”‚     616.59 / 619.67 ยฑ1.77 / 621.62 ms โ”‚     no change โ”‚
   โ”‚ QQuery 17 โ”‚     614.39 / 624.14 ยฑ7.71 / 632.38 ms โ”‚     616.16 / 623.59 ยฑ5.43 / 630.17 ms โ”‚     no change โ”‚
   โ”‚ QQuery 18 โ”‚ 1253.14 / 1267.36 ยฑ11.00 / 1280.60 ms โ”‚  1248.19 / 1257.72 ยฑ7.72 / 1265.73 ms โ”‚     no change โ”‚
   โ”‚ QQuery 19 โ”‚        29.05 / 29.35 ยฑ0.31 / 29.95 ms โ”‚        28.66 / 29.30 ยฑ0.58 / 30.18 ms โ”‚     no change โ”‚
   โ”‚ QQuery 20 โ”‚    517.45 / 528.92 ยฑ18.95 / 566.76 ms โ”‚     525.24 / 529.80 ยฑ3.91 / 535.17 ms โ”‚     no change โ”‚
   โ”‚ QQuery 21 โ”‚     600.89 / 607.92 ยฑ4.67 / 614.88 ms โ”‚     606.00 / 610.39 ยฑ3.30 / 614.47 ms โ”‚     no change โ”‚
   โ”‚ QQuery 22 โ”‚  1070.64 / 1079.02 ยฑ8.00 / 1089.22 ms โ”‚ 1078.15 / 1095.45 ยฑ12.80 / 1117.46 ms โ”‚     no change โ”‚
   โ”‚ QQuery 23 โ”‚ 3360.61 / 3401.23 ยฑ30.44 / 3447.41 ms โ”‚ 3335.05 / 3372.58 ยฑ22.28 / 3398.53 ms โ”‚     no change โ”‚
   โ”‚ QQuery 24 โ”‚        41.99 / 42.57 ยฑ0.46 / 43.26 ms โ”‚        41.68 / 42.45 ยฑ0.54 / 43.06 ms โ”‚     no change โ”‚
   โ”‚ QQuery 25 โ”‚     115.06 / 120.06 ยฑ7.13 / 133.94 ms โ”‚     114.06 / 114.91 ยฑ0.87 / 116.59 ms โ”‚     no change โ”‚
   โ”‚ QQuery 26 โ”‚        42.64 / 43.22 ยฑ0.68 / 44.52 ms โ”‚        42.63 / 44.51 ยฑ1.15 / 46.21 ms โ”‚     no change โ”‚
   โ”‚ QQuery 27 โ”‚     670.19 / 673.51 ยฑ2.53 / 676.71 ms โ”‚     665.22 / 674.44 ยฑ6.83 / 685.39 ms โ”‚     no change โ”‚
   โ”‚ QQuery 28 โ”‚  3020.20 / 3032.53 ยฑ7.25 / 3042.60 ms โ”‚  3034.07 / 3043.92 ยฑ5.02 / 3047.17 ms โ”‚     no change โ”‚
   โ”‚ QQuery 29 โ”‚        42.92 / 49.87 ยฑ8.68 / 64.44 ms โ”‚        42.72 / 51.32 ยฑ8.44 / 66.05 ms โ”‚     no change โ”‚
   โ”‚ QQuery 30 โ”‚     310.59 / 314.71 ยฑ3.56 / 320.88 ms โ”‚     309.32 / 315.80 ยฑ6.21 / 326.49 ms โ”‚     no change โ”‚
   โ”‚ QQuery 31 โ”‚     296.12 / 303.16 ยฑ4.85 / 308.13 ms โ”‚     300.68 / 309.74 ยฑ9.96 / 328.65 ms โ”‚     no change โ”‚
   โ”‚ QQuery 32 โ”‚   999.82 / 1005.51 ยฑ3.55 / 1010.32 ms โ”‚   990.79 / 1004.42 ยฑ9.05 / 1015.78 ms โ”‚     no change โ”‚
   โ”‚ QQuery 33 โ”‚ 1432.71 / 1457.69 ยฑ13.63 / 1471.29 ms โ”‚ 1451.44 / 1470.33 ยฑ20.03 / 1507.76 ms โ”‚     no change โ”‚
   โ”‚ QQuery 34 โ”‚ 1460.74 / 1484.06 ยฑ16.42 / 1503.70 ms โ”‚ 1472.65 / 1482.70 ยฑ11.30 / 1503.95 ms โ”‚     no change โ”‚
   โ”‚ QQuery 35 โ”‚    289.92 / 320.76 ยฑ34.80 / 377.25 ms โ”‚    238.14 / 255.97 ยฑ27.33 / 309.81 ms โ”‚ +1.25x faster โ”‚
   โ”‚ QQuery 36 โ”‚        63.06 / 65.39 ยฑ2.03 / 68.90 ms โ”‚        63.66 / 70.86 ยฑ6.09 / 81.47 ms โ”‚  1.08x slower โ”‚
   โ”‚ QQuery 37 โ”‚        36.21 / 38.40 ยฑ3.99 / 46.37 ms โ”‚        35.95 / 38.10 ยฑ3.42 / 44.87 ms โ”‚     no change โ”‚
   โ”‚ QQuery 38 โ”‚        41.53 / 46.49 ยฑ4.03 / 53.62 ms โ”‚        41.39 / 44.42 ยฑ1.81 / 46.44 ms โ”‚     no change โ”‚
   โ”‚ QQuery 39 โ”‚     128.58 / 134.91 ยฑ5.38 / 142.50 ms โ”‚     129.60 / 137.12 ยฑ6.81 / 149.96 ms โ”‚     no change โ”‚
   โ”‚ QQuery 40 โ”‚        15.04 / 16.71 ยฑ2.36 / 21.33 ms โ”‚        14.66 / 18.79 ยฑ4.54 / 25.14 ms โ”‚  1.12x slower โ”‚
   โ”‚ QQuery 41 โ”‚        14.48 / 16.66 ยฑ4.06 / 24.78 ms โ”‚        14.10 / 14.79 ยฑ0.68 / 16.03 ms โ”‚ +1.13x faster โ”‚
   โ”‚ QQuery 42 โ”‚        13.81 / 14.12 ยฑ0.27 / 14.57 ms โ”‚        13.61 / 13.84 ยฑ0.24 / 14.20 ms โ”‚     no change โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   โ”ƒ Benchmark Summary                                 โ”ƒ            โ”ƒ
   โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   โ”‚ Total Time (HEAD)                                 โ”‚ 20188.14ms โ”‚
   โ”‚ Total Time (speed-up-single-primitive-group-by)   โ”‚ 20041.89ms โ”‚
   โ”‚ Average Time (HEAD)                               โ”‚   469.49ms โ”‚
   โ”‚ Average Time (speed-up-single-primitive-group-by) โ”‚   466.09ms โ”‚
   โ”‚ Queries Faster                                    โ”‚          5 โ”‚
   โ”‚ Queries Slower                                    โ”‚          2 โ”‚
   โ”‚ Queries with No Change                            โ”‚         36 โ”‚
   โ”‚ Queries with Failure                              โ”‚          0 โ”‚
   โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   ```
   
   </p>
   </details>
   
   <details><summary>Resource Usage</summary>
   
   **clickbench_partitioned โ€” base (merge-base)**
   | Metric | Value |
   |--------|-------|
   | Wall time | 105.0s |
   | Peak memory | 31.4 GiB |
   | Avg memory | 23.4 GiB |
   | CPU user | 1065.7s |
   | CPU sys | 65.1s |
   | Peak spill | 0 B |
   
   **clickbench_partitioned โ€” branch**
   | Metric | Value |
   |--------|-------|
   | Wall time | 105.0s |
   | Peak memory | 30.6 GiB |
   | Avg memory | 23.1 GiB |
   | CPU user | 1056.4s |
   | CPU sys | 65.6s |
   | 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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4363668679

   run benchmark clickbench_partitioned


-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364094378

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364085635)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4364085635-1977-z7vq2 6.12.68+ #1 SMP Wed Apr  1 02:23:28 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (08629a74dc3ee2842f44e43fe8ee96f8c3af7dea) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..08629a74dc3ee2842f44e43fe8ee96f8c3af7dea) using: tpch
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364091985

   ๐Ÿค– Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21977#issuecomment-4364085635)
   **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4364085635-1976-zl9fc 6.12.68+ #1 SMP Wed Apr  1 02:23:28 UTC 2026 aarch64 GNU/Linux`
   <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>
   
   Comparing speed-up-single-primitive-group-by (08629a74dc3ee2842f44e43fe8ee96f8c3af7dea) to 948cd09 (merge-base) [diff](https://github.com/apache/datafusion/compare/948cd09c10604561a0f755eebf8c25b27a79250f..08629a74dc3ee2842f44e43fe8ee96f8c3af7dea) using: tpcds
   Results will be posted here when complete
   
   ---
   [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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning, use less memory [datafusion]

Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4364085635

   run benchmarks


-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org


Re: [PR] perf: speed up primitive group value interning [datafusion]

Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21977:
URL: https://github.com/apache/datafusion/pull/21977#issuecomment-4360963663

   run benchmarks


-- 
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: github-unsubscribe@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org