Skip to content

[do not merge] Enable dynamic task count by default - #642

Closed
gabotechs wants to merge 1 commit into
mainfrom
gabrielmusat/enable-aqe-by-default
Closed

[do not merge] Enable dynamic task count by default#642
gabotechs wants to merge 1 commit into
mainfrom
gabrielmusat/enable-aqe-by-default

Conversation

@gabotechs

Copy link
Copy Markdown
Collaborator

Do not merge!

@gabotechs

gabotechs commented Aug 13, 2026

Copy link
Copy Markdown
Collaborator Author

benchmarks run tpch/sf1 tpch/sf10 tpch/sf100 tpcds/sf1 clickbench/0-100

@gabot-0

gabot-0 commented Aug 13, 2026

Copy link
Copy Markdown

Requested by this comment.

Benchmark results

Compared: Base 81b887eb0129PR head bc15154d43cf · View exact source diff

=== Comparing tpch/sf1 results 'datafusion-benchmark-base' [prev] with 'datafusion-benchmark-head' [new] ===
TOTAL: prev=10852 ms, new=11115 ms, diff=1.02 slower ✖
Show full query output
      q1: prev= 360 ms, new= 434 ms, diff=1.21 slower ✖
      q2: prev= 668 ms, new= 601 ms, diff=1.11 faster ✔
      q3: prev= 480 ms, new= 622 ms, diff=1.30 slower ✖
      q4: prev= 327 ms, new= 313 ms, diff=1.04 faster ✔
      q5: prev= 558 ms, new= 618 ms, diff=1.11 slower ✖
      q6: prev= 148 ms, new= 157 ms, diff=1.06 slower ✖
      q7: prev= 809 ms, new= 517 ms, diff=1.56 faster ✅
      q8: prev= 843 ms, new= 840 ms, diff=1.00 faster ✔
      q9: prev= 868 ms, new= 922 ms, diff=1.06 slower ✖
     q10: prev= 588 ms, new= 555 ms, diff=1.06 faster ✔
     q11: prev= 392 ms, new= 349 ms, diff=1.12 faster ✔
     q12: prev= 335 ms, new= 476 ms, diff=1.42 slower ✖
     q13: prev= 324 ms, new= 315 ms, diff=1.03 faster ✔
     q14: prev= 337 ms, new= 270 ms, diff=1.25 faster ✔
     q15: prev= 352 ms, new= 359 ms, diff=1.02 slower ✖
     q16: prev= 186 ms, new= 189 ms, diff=1.02 slower ✖
     q17: prev= 542 ms, new= 550 ms, diff=1.01 slower ✖
     q18: prev= 740 ms, new= 618 ms, diff=1.20 faster ✔
     q19: prev= 425 ms, new= 349 ms, diff=1.22 faster ✔
     q20: prev= 575 ms, new= 525 ms, diff=1.10 faster ✔
     q21: prev= 731 ms, new=1233 ms, diff=1.69 slower ❌
     q22: prev= 264 ms, new= 303 ms, diff=1.15 slower ✖
=== Comparing tpch/sf10 results 'datafusion-benchmark-base' [prev] with 'datafusion-benchmark-head' [new] ===
TOTAL: prev=23135 ms, new=22798 ms, diff=1.01 faster ✔
Show full query output
      q1: prev=1026 ms, new= 511 ms, diff=2.01 faster ✅
      q2: prev=1182 ms, new= 954 ms, diff=1.24 faster ✔
      q3: prev=1289 ms, new= 980 ms, diff=1.32 faster ✔
      q4: prev= 627 ms, new= 665 ms, diff=1.06 slower ✖
      q5: prev= 966 ms, new=1138 ms, diff=1.18 slower ✖
      q6: prev= 430 ms, new= 871 ms, diff=2.03 slower ❌
      q7: prev=1388 ms, new=1392 ms, diff=1.00 slower ✖
      q8: prev=2010 ms, new=1838 ms, diff=1.09 faster ✔
      q9: prev=2068 ms, new=2144 ms, diff=1.04 slower ✖
     q10: prev=1343 ms, new=1457 ms, diff=1.08 slower ✖
     q11: prev= 729 ms, new=1031 ms, diff=1.41 slower ✖
     q12: prev= 689 ms, new=1132 ms, diff=1.64 slower ❌
     q13: prev= 762 ms, new= 726 ms, diff=1.05 faster ✔
     q14: prev= 587 ms, new= 880 ms, diff=1.50 slower ✖
     q15: prev=1179 ms, new=1387 ms, diff=1.18 slower ✖
     q16: prev= 409 ms, new= 410 ms, diff=1.00 slower ✖
     q17: prev= 928 ms, new= 904 ms, diff=1.03 faster ✔
     q18: prev=1723 ms, new=1459 ms, diff=1.18 faster ✔
     q19: prev= 581 ms, new= 514 ms, diff=1.13 faster ✔
     q20: prev=1478 ms, new= 734 ms, diff=2.01 faster ✅
     q21: prev=1404 ms, new=1311 ms, diff=1.07 faster ✔
     q22: prev= 337 ms, new= 360 ms, diff=1.07 slower ✖
=== Comparing tpch/sf100 results 'datafusion-benchmark-base' [prev] with 'datafusion-benchmark-head' [new] ===
TOTAL: prev=100556 ms, new=77664 ms, diff=1.29 faster ✅
Show full query output
      q1: prev=4079 ms, new=3494 ms, diff=1.17 faster ✔
      q2: prev=1976 ms, new=2098 ms, diff=1.06 slower ✖
      q3: prev=3706 ms, new=2771 ms, diff=1.34 faster ✔
      q4: prev=1772 ms, new=1464 ms, diff=1.21 faster ✔
      q5: prev=4604 ms, new=4038 ms, diff=1.14 faster ✔
      q6: prev=1360 ms, new=1389 ms, diff=1.02 slower ✖
      q7: prev=4805 ms, new=4591 ms, diff=1.05 faster ✔
      q8: prev=4991 ms, new=4957 ms, diff=1.01 faster ✔
      q9: prev=6307 ms, new=6541 ms, diff=1.04 slower ✖
     q10: prev=6136 ms, new=6874 ms, diff=1.12 slower ✖
     q11: prev=1232 ms, new=1292 ms, diff=1.05 slower ✖
     q12: prev=2026 ms, new=1968 ms, diff=1.03 faster ✔
     q13: prev=1897 ms, new=1913 ms, diff=1.01 slower ✖
     q14: prev=1924 ms, new=1871 ms, diff=1.03 faster ✔
     q15: prev=3740 ms, new=3725 ms, diff=1.00 faster ✔
     q16: prev= 897 ms, new= 776 ms, diff=1.16 faster ✔
     q17: prev=5162 ms, new=4640 ms, diff=1.11 faster ✔
     q18: prev=6080 ms, new=5720 ms, diff=1.06 faster ✔
     q19: prev=8827 ms, new=3885 ms, diff=2.27 faster ✅
     q20: prev=20004 ms, new=5436 ms, diff=3.68 faster ✅
     q21: prev=7805 ms, new=7158 ms, diff=1.09 faster ✔
     q22: prev=1226 ms, new=1063 ms, diff=1.15 faster ✔
=== Comparing tpcds/sf1 results 'datafusion-benchmark-base' [prev] with 'datafusion-benchmark-head' [new] ===
TOTAL: prev=62401 ms, new=62704 ms, diff=1.00 slower ✖
Show full query output
      q1: prev= 355 ms, new= 300 ms, diff=1.18 faster ✔
      q2: prev= 681 ms, new= 729 ms, diff=1.07 slower ✖
      q3: prev= 381 ms, new= 515 ms, diff=1.35 slower ✖
      q4: prev=2996 ms, new=3008 ms, diff=1.00 slower ✖
      q5: prev= 743 ms, new= 825 ms, diff=1.11 slower ✖
      q6: prev= 703 ms, new= 757 ms, diff=1.08 slower ✖
      q7: prev= 337 ms, new= 504 ms, diff=1.50 slower ✖
      q8: prev= 332 ms, new= 288 ms, diff=1.15 faster ✔
      q9: prev= 339 ms, new= 506 ms, diff=1.49 slower ✖
     q10: prev= 727 ms, new= 582 ms, diff=1.25 faster ✔
     q11: prev=1989 ms, new=1891 ms, diff=1.05 faster ✔
     q12: prev= 136 ms, new= 133 ms, diff=1.02 faster ✔
     q13: prev=1053 ms, new=1142 ms, diff=1.08 slower ✖
     q14: prev=1229 ms, new=1434 ms, diff=1.17 slower ✖
     q15: prev= 403 ms, new= 351 ms, diff=1.15 faster ✔
     q16: prev= 584 ms, new= 334 ms, diff=1.75 faster ✅
     q17: prev= 813 ms, new= 596 ms, diff=1.36 faster ✔
     q18: prev= 542 ms, new= 591 ms, diff=1.09 slower ✖
     q19: prev= 394 ms, new= 467 ms, diff=1.19 slower ✖
     q20: prev= 153 ms, new= 136 ms, diff=1.13 faster ✔
     q21: prev= 173 ms, new= 174 ms, diff=1.01 slower ✖
     q22: prev= 521 ms, new= 487 ms, diff=1.07 faster ✔
     q23: prev=1034 ms, new=1084 ms, diff=1.05 slower ✖
     q24: prev=1248 ms, new=1366 ms, diff=1.09 slower ✖
     q25: prev= 441 ms, new= 612 ms, diff=1.39 slower ✖
     q26: prev= 239 ms, new= 248 ms, diff=1.04 slower ✖
     q27: prev= 403 ms, new= 504 ms, diff=1.25 slower ✖
     q28: prev= 197 ms, new= 316 ms, diff=1.60 slower ❌
     q29: prev= 415 ms, new= 428 ms, diff=1.03 slower ✖
q30: Previously failed, and now also failed ❌
     q31: prev=1178 ms, new=1124 ms, diff=1.05 faster ✔
     q32: prev= 216 ms, new= 169 ms, diff=1.28 faster ✔
     q33: prev= 351 ms, new= 321 ms, diff=1.09 faster ✔
     q34: prev= 278 ms, new= 208 ms, diff=1.34 faster ✔
     q35: prev= 708 ms, new= 618 ms, diff=1.15 faster ✔
     q36: prev= 294 ms, new= 252 ms, diff=1.17 faster ✔
     q37: prev= 428 ms, new= 329 ms, diff=1.30 faster ✔
     q38: prev= 550 ms, new= 423 ms, diff=1.30 faster ✔
     q39: prev= 380 ms, new= 310 ms, diff=1.23 faster ✔
     q40: prev= 357 ms, new= 415 ms, diff=1.16 slower ✖
     q41: prev=  84 ms, new= 131 ms, diff=1.56 slower ❌
     q42: prev= 196 ms, new= 236 ms, diff=1.20 slower ✖
     q43: prev= 146 ms, new= 127 ms, diff=1.15 faster ✔
     q44: prev= 374 ms, new= 246 ms, diff=1.52 faster ✅
     q45: prev= 425 ms, new= 357 ms, diff=1.19 faster ✔
     q46: prev= 509 ms, new= 414 ms, diff=1.23 faster ✔
     q47: prev= 680 ms, new= 792 ms, diff=1.16 slower ✖
     q48: prev= 541 ms, new= 673 ms, diff=1.24 slower ✖
     q49: prev= 585 ms, new= 551 ms, diff=1.06 faster ✔
     q50: prev= 445 ms, new= 462 ms, diff=1.04 slower ✖
     q51: prev= 412 ms, new= 479 ms, diff=1.16 slower ✖
     q52: prev= 189 ms, new= 288 ms, diff=1.52 slower ❌
     q53: prev= 161 ms, new= 165 ms, diff=1.02 slower ✖
     q54: prev= 544 ms, new= 741 ms, diff=1.36 slower ✖
     q55: prev= 112 ms, new= 159 ms, diff=1.42 slower ✖
     q56: prev= 280 ms, new= 291 ms, diff=1.04 slower ✖
     q57: prev= 475 ms, new= 464 ms, diff=1.02 faster ✔
     q58: prev= 636 ms, new= 608 ms, diff=1.05 faster ✔
     q59: prev= 284 ms, new= 310 ms, diff=1.09 slower ✖
     q60: prev= 313 ms, new= 346 ms, diff=1.11 slower ✖
     q61: prev=1058 ms, new= 897 ms, diff=1.18 faster ✔
     q62: prev= 399 ms, new= 429 ms, diff=1.08 slower ✖
     q63: prev= 178 ms, new= 189 ms, diff=1.06 slower ✖
     q64: prev=3019 ms, new=3089 ms, diff=1.02 slower ✖
     q65: prev= 493 ms, new= 351 ms, diff=1.40 faster ✔
     q66: prev= 482 ms, new= 429 ms, diff=1.12 faster ✔
     q67: prev= 538 ms, new= 554 ms, diff=1.03 slower ✖
     q68: prev= 524 ms, new= 435 ms, diff=1.20 faster ✔
     q69: prev= 651 ms, new= 686 ms, diff=1.05 slower ✖
     q70: prev= 450 ms, new= 392 ms, diff=1.15 faster ✔
     q71: prev= 228 ms, new= 201 ms, diff=1.13 faster ✔
     q72: prev=11726 ms, new=11504 ms, diff=1.02 faster ✔
     q73: prev= 281 ms, new= 231 ms, diff=1.22 faster ✔
     q74: prev= 800 ms, new= 920 ms, diff=1.15 slower ✖
     q75: prev=1007 ms, new=1008 ms, diff=1.00 slower ✖
     q76: prev= 409 ms, new= 240 ms, diff=1.70 faster ✅
     q77: prev= 491 ms, new= 476 ms, diff=1.03 faster ✔
     q78: prev= 961 ms, new= 850 ms, diff=1.13 faster ✔
     q79: prev= 241 ms, new= 236 ms, diff=1.02 faster ✔
     q80: prev= 617 ms, new= 667 ms, diff=1.08 slower ✖
     q81: prev= 293 ms, new= 276 ms, diff=1.06 faster ✔
     q82: prev= 303 ms, new= 316 ms, diff=1.04 slower ✖
     q83: prev= 265 ms, new= 303 ms, diff=1.14 slower ✖
     q84: prev= 314 ms, new= 312 ms, diff=1.01 faster ✔
     q85: prev= 512 ms, new= 533 ms, diff=1.04 slower ✖
     q86: prev= 154 ms, new= 157 ms, diff=1.02 slower ✖
     q87: prev= 455 ms, new= 423 ms, diff=1.08 faster ✔
     q88: prev= 437 ms, new= 415 ms, diff=1.05 faster ✔
     q89: prev= 203 ms, new= 238 ms, diff=1.17 slower ✖
     q90: prev= 187 ms, new= 146 ms, diff=1.28 faster ✔
     q91: prev= 402 ms, new= 457 ms, diff=1.14 slower ✖
     q92: prev= 241 ms, new= 272 ms, diff=1.13 slower ✖
     q93: prev= 266 ms, new= 262 ms, diff=1.02 faster ✔
     q94: prev= 319 ms, new= 430 ms, diff=1.35 slower ✖
     q95: prev= 516 ms, new= 670 ms, diff=1.30 slower ✖
     q96: prev= 227 ms, new= 252 ms, diff=1.11 slower ✖
     q97: prev= 280 ms, new= 297 ms, diff=1.06 slower ✖
     q98: prev= 174 ms, new= 228 ms, diff=1.31 slower ✖
     q99: prev= 608 ms, new= 616 ms, diff=1.01 slower ✖
=== Comparing clickbench/0-100 results 'datafusion-benchmark-base' [prev] with 'datafusion-benchmark-head' [new] ===
TOTAL: prev=32797 ms, new=30156 ms, diff=1.09 faster ✔
Show full query output
      q0: prev=   2 ms, new=   2 ms, diff=1.00 slower ✖
      q1: prev= 324 ms, new= 337 ms, diff=1.04 slower ✖
      q2: prev= 365 ms, new= 298 ms, diff=1.22 faster ✔
      q3: prev= 461 ms, new= 393 ms, diff=1.17 faster ✔
      q4: prev= 460 ms, new= 406 ms, diff=1.13 faster ✔
      q5: prev= 534 ms, new= 542 ms, diff=1.01 slower ✖
      q6: prev=   2 ms, new=   2 ms, diff=1.00 slower ✖
      q7: prev= 354 ms, new= 274 ms, diff=1.29 faster ✔
      q8: prev= 510 ms, new= 510 ms, diff=1.00 slower ✖
      q9: prev= 546 ms, new= 587 ms, diff=1.08 slower ✖
     q10: prev= 490 ms, new= 433 ms, diff=1.13 faster ✔
     q11: prev= 478 ms, new= 390 ms, diff=1.23 faster ✔
     q12: prev= 518 ms, new= 528 ms, diff=1.02 slower ✖
     q13: prev= 717 ms, new= 704 ms, diff=1.02 faster ✔
     q14: prev= 518 ms, new= 571 ms, diff=1.10 slower ✖
     q15: prev= 410 ms, new= 380 ms, diff=1.08 faster ✔
     q16: prev= 692 ms, new= 681 ms, diff=1.02 faster ✔
     q17: prev= 659 ms, new= 657 ms, diff=1.00 faster ✔
     q18: prev=1035 ms, new= 962 ms, diff=1.08 faster ✔
     q19: prev= 288 ms, new= 341 ms, diff=1.18 slower ✖
     q20: prev=1479 ms, new=1417 ms, diff=1.04 faster ✔
     q21: prev=1461 ms, new=1511 ms, diff=1.03 slower ✖
     q22: prev=2236 ms, new=1872 ms, diff=1.19 faster ✔
     q23: prev=5379 ms, new=4626 ms, diff=1.16 faster ✔
     q24: prev= 522 ms, new= 531 ms, diff=1.02 slower ✖
     q25: prev= 424 ms, new= 336 ms, diff=1.26 faster ✔
     q26: prev= 993 ms, new= 444 ms, diff=2.24 faster ✅
     q27: prev=1651 ms, new=1541 ms, diff=1.07 faster ✔
     q28: prev=2517 ms, new=2346 ms, diff=1.07 faster ✔
     q29: prev= 333 ms, new= 267 ms, diff=1.25 faster ✔
     q30: prev= 725 ms, new= 503 ms, diff=1.44 faster ✔
     q31: prev= 764 ms, new= 896 ms, diff=1.17 slower ✖
     q32: prev= 784 ms, new= 721 ms, diff=1.09 faster ✔
     q33: prev=1892 ms, new=1864 ms, diff=1.02 faster ✔
     q34: prev=1868 ms, new=1900 ms, diff=1.02 slower ✖
     q35: prev= 406 ms, new= 383 ms, diff=1.06 faster ✔
q36: Previously failed, and now also failed ❌
q37: Previously failed, and now also failed ❌
q38: Previously failed, and now also failed ❌
q39: Previously failed, and now also failed ❌
q40: Previously failed, and now also failed ❌
q41: Previously failed, and now also failed ❌
q42: Previously failed, and now also failed ❌
Verification and run details

Job 37 captured both immutable revisions when the request was queued. The bot fetched and checked out each full commit SHA in detached HEAD, then built and deployed the datafusion-distributed-benchmarks --bin worker target from that checkout.

Identity Base PR head
Source commit 81b887eb012980ad32a9de9da17b8109aedb49ee bc15154d43cf715db887a93aa942a63277416f0a
Phase Base PR head
Build and deployment 2m 15s 1m 26s
All benchmarks 24m 37s 22m 6s
Benchmark tpch/sf1 1m 20s 1m 16s
Benchmark tpch/sf10 2m 42s 2m 32s
Benchmark tpch/sf100 10m 18s 8m 22s
Benchmark tpcds/sf1 6m 39s 6m 38s
Benchmark clickbench/0-100 3m 38s 3m 19s

Workload: tpch/sf1, tpch/sf10, tpch/sf100, tpcds/sf1, clickbench/0-100 · all queries · 1 warmup + 5 measured iterations per query

Capacity: 12 c5n.2xlarge nodes for both revisions

Other timings: Queue 1s · Dataset validation 2s · Total 50m 39s

@gabotechs

Copy link
Copy Markdown
Collaborator Author

benchmarks run tpch/sf100

@gabot-0

gabot-0 commented Aug 14, 2026

Copy link
Copy Markdown

Requested by this comment.

Benchmark results

Compared: Base 81b887eb0129PR head bc15154d43cf · View exact source diff

=== Comparing tpch/sf100 results 'datafusion-benchmark-base' [prev] with 'datafusion-benchmark-head' [new] ===
TASKS: prev=946.0, new=1086.0, diff=140.0 more (14.8%) (sum of per-query averages)
TOTAL: prev=122355 ms, new=96138 ms, diff=1.27 faster ✅
Show full query output
      q1: prev=2770 ms, new=2792 ms, diff=1.01 slower ✖, tasks: prev=18.0, new=13.0, diff=5.0 fewer (27.8%)
      q2: prev=3098 ms, new=2444 ms, diff=1.27 faster ✔, tasks: prev=40.0, new=74.0, diff=34.0 more (85.0%)
      q3: prev=2878 ms, new=2893 ms, diff=1.01 slower ✖, tasks: prev=48.0, new=56.0, diff=8.0 more (16.7%)
      q4: prev=1318 ms, new=1377 ms, diff=1.04 slower ✖, tasks: prev=38.0, new=37.0, diff=1.0 fewer (2.6%)
      q5: prev=4540 ms, new=4297 ms, diff=1.06 faster ✔, tasks: prev=52.0, new=60.0, diff=8.0 more (15.4%)
      q6: prev=1438 ms, new=1397 ms, diff=1.03 faster ✔, tasks: prev=12.0, new=12.0, diff=no change
      q7: prev=9240 ms, new=8750 ms, diff=1.06 faster ✔, tasks: prev=53.0, new=61.0, diff=8.0 more (15.1%)
      q8: prev=9151 ms, new=4788 ms, diff=1.91 faster ✅, tasks: prev=73.0, new=74.0, diff=1.0 more (1.4%)
      q9: prev=12383 ms, new=9309 ms, diff=1.33 faster ✔, tasks: prev=77.0, new=87.0, diff=10.0 more (13.0%)
     q10: prev=12587 ms, new=8752 ms, diff=1.44 faster ✔, tasks: prev=40.0, new=61.0, diff=21.0 more (52.5%)
     q11: prev=6066 ms, new=3243 ms, diff=1.87 faster ✅, tasks: prev=24.0, new=31.0, diff=7.0 more (29.2%)
     q12: prev=6161 ms, new=1792 ms, diff=3.44 faster ✅, tasks: prev=44.0, new=37.0, diff=7.0 fewer (15.9%)
     q13: prev=3888 ms, new=1625 ms, diff=2.39 faster ✅, tasks: prev=32.0, new=33.0, diff=1.0 more (3.1%)
     q14: prev=3572 ms, new=1831 ms, diff=1.95 faster ✅, tasks: prev=24.0, new=29.0, diff=5.0 more (20.8%)
     q15: prev=4028 ms, new=3518 ms, diff=1.14 faster ✔, tasks: prev=30.0, new=39.0, diff=9.0 more (30.0%)
     q16: prev= 808 ms, new= 526 ms, diff=1.54 faster ✅, tasks: prev=47.0, new=57.0, diff=10.0 more (21.3%)
     q17: prev=5043 ms, new=4771 ms, diff=1.06 faster ✔, tasks: prev=40.0, new=41.0, diff=1.0 more (2.5%)
     q18: prev=5642 ms, new=5168 ms, diff=1.09 faster ✔, tasks: prev=68.0, new=69.0, diff=1.0 more (1.5%)
     q19: prev=2451 ms, new=1752 ms, diff=1.40 faster ✔, tasks: prev=24.0, new=36.0, diff=12.0 more (50.0%)
     q20: prev=2798 ms, new=2218 ms, diff=1.26 faster ✔, tasks: prev=65.0, new=63.0, diff=2.0 fewer (3.1%)
     q21: prev=21683 ms, new=22077 ms, diff=1.02 slower ✖, tasks: prev=57.0, new=66.0, diff=9.0 more (15.8%)
     q22: prev= 812 ms, new= 818 ms, diff=1.01 slower ✖, tasks: prev=40.0, new=50.0, diff=10.0 more (25.0%)
Verification and run details

Job 38 captured both immutable revisions when the request was queued. The bot fetched and checked out each full commit SHA in detached HEAD, then built and deployed the datafusion-distributed-benchmarks --bin worker target from that checkout.

Identity Base PR head
Source commit 81b887eb012980ad32a9de9da17b8109aedb49ee bc15154d43cf715db887a93aa942a63277416f0a
Phase Base PR head
Build and deployment 1m 26s 1m 25s
All benchmarks 12m 19s 10m 0s
Benchmark tpch/sf100 12m 19s 10m 0s

Workload: tpch/sf100 · all queries · 1 warmup + 5 measured iterations per query

Capacity: 12 c5n.2xlarge nodes for both revisions

Other timings: Queue 1s · Dataset validation 0s · Total 25m 17s

@gabotechs

Copy link
Copy Markdown
Collaborator Author

benchmarks run tpcds/sf1

@gabot-0

gabot-0 commented Aug 14, 2026

Copy link
Copy Markdown

Requested by this comment.

Benchmark results

Compared: Base 81b887eb0129PR head bc15154d43cf · View exact source diff

=== Comparing tpcds/sf1 results 'datafusion-benchmark-base' [prev] with 'datafusion-benchmark-head' [new] ===
TASKS: prev=261.0, new=1376.4, diff=1115.4 more (427.4%) (sum of per-query averages)
TOTAL: prev=75546 ms, new=68531 ms, diff=1.10 faster ✅
Show full query output
      q1: prev= 300 ms, new= 237 ms, diff=1.27 faster ✔, tasks: prev=0.0, new=9.0, diff=9.0 more
      q2: prev= 298 ms, new= 234 ms, diff=1.27 faster ✔, tasks: prev=0.0, new=9.0, diff=9.0 more
      q3: prev= 452 ms, new= 372 ms, diff=1.22 faster ✔, tasks: prev=2.0, new=6.0, diff=4.0 more (200.0%)
      q4: prev=2142 ms, new=1280 ms, diff=1.67 faster ✅, tasks: prev=0.0, new=18.0, diff=18.0 more
      q5: prev= 806 ms, new= 255 ms, diff=3.16 faster ✅, tasks: prev=5.0, new=13.0, diff=8.0 more (160.0%)
      q6: prev=6659 ms, new=7062 ms, diff=1.06 slower ✖, tasks: prev=0.0, new=8.0, diff=8.0 more
      q7: prev= 332 ms, new= 245 ms, diff=1.36 faster ✔, tasks: prev=2.0, new=10.0, diff=8.0 more (400.0%)
      q8: prev= 273 ms, new= 319 ms, diff=1.17 slower ✖, tasks: prev=0.0, new=9.0, diff=9.0 more
      q9: prev= 261 ms, new= 341 ms, diff=1.31 slower ✖, tasks: prev=30.0, new=34.0, diff=4.0 more (13.3%)
     q10: prev= 506 ms, new= 273 ms, diff=1.85 faster ✅, tasks: prev=0.0, new=13.0, diff=13.0 more
     q11: prev=1449 ms, new= 881 ms, diff=1.64 faster ✅, tasks: prev=0.0, new=12.0, diff=12.0 more
     q12: prev= 139 ms, new= 155 ms, diff=1.12 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
     q13: prev= 509 ms, new= 552 ms, diff=1.08 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
     q14: prev= 743 ms, new=1155 ms, diff=1.55 slower ❌, tasks: prev=9.0, new=58.6, diff=49.6 more (551.1%)
     q15: prev= 253 ms, new= 192 ms, diff=1.32 faster ✔, tasks: prev=2.0, new=6.0, diff=4.0 more (200.0%)
     q16: prev= 361 ms, new= 255 ms, diff=1.42 faster ✔, tasks: prev=10.0, new=10.8, diff=0.8 more (8.0%)
     q17: prev= 388 ms, new= 281 ms, diff=1.38 faster ✔, tasks: prev=7.0, new=16.4, diff=9.4 more (134.3%)
     q18: prev= 499 ms, new= 415 ms, diff=1.20 faster ✔, tasks: prev=2.0, new=12.0, diff=10.0 more (500.0%)
     q19: prev= 245 ms, new= 274 ms, diff=1.12 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
     q20: prev= 139 ms, new= 174 ms, diff=1.25 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
     q21: prev= 149 ms, new= 266 ms, diff=1.79 slower ❌, tasks: prev=0.0, new=4.0, diff=4.0 more
     q22: prev= 559 ms, new= 584 ms, diff=1.04 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
     q23: prev= 780 ms, new= 714 ms, diff=1.09 faster ✔, tasks: prev=8.0, new=38.0, diff=30.0 more (375.0%)
     q24: prev= 489 ms, new= 422 ms, diff=1.16 faster ✔, tasks: prev=4.0, new=21.0, diff=17.0 more (425.0%)
     q25: prev= 412 ms, new= 259 ms, diff=1.59 faster ✅, tasks: prev=7.0, new=15.4, diff=8.4 more (120.0%)
     q26: prev= 253 ms, new= 162 ms, diff=1.56 faster ✅, tasks: prev=2.0, new=8.0, diff=6.0 more (300.0%)
     q27: prev= 871 ms, new= 354 ms, diff=2.46 faster ✅, tasks: prev=9.0, new=30.0, diff=21.0 more (233.3%)
     q28: prev= 146 ms, new= 197 ms, diff=1.35 slower ✖, tasks: prev=12.0, new=72.0, diff=60.0 more (500.0%)
     q29: prev= 416 ms, new= 275 ms, diff=1.51 faster ✅, tasks: prev=7.0, new=16.4, diff=9.4 more (134.3%)
q30: Previously failed, and now also failed ❌
     q31: prev= 786 ms, new= 279 ms, diff=2.82 faster ✅, tasks: prev=6.0, new=33.0, diff=27.0 more (450.0%)
     q32: prev= 169 ms, new= 192 ms, diff=1.14 slower ✖, tasks: prev=0.0, new=6.0, diff=6.0 more
     q33: prev= 218 ms, new= 292 ms, diff=1.34 slower ✖, tasks: prev=0.0, new=14.0, diff=14.0 more
     q34: prev= 236 ms, new= 393 ms, diff=1.67 slower ❌, tasks: prev=0.0, new=6.0, diff=6.0 more
     q35: prev= 520 ms, new= 375 ms, diff=1.39 faster ✔, tasks: prev=0.0, new=12.0, diff=12.0 more
     q36: prev= 343 ms, new= 328 ms, diff=1.05 faster ✔, tasks: prev=4.0, new=17.0, diff=13.0 more (325.0%)
     q37: prev= 318 ms, new= 318 ms, diff=1.00 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
     q38: prev= 331 ms, new= 214 ms, diff=1.55 faster ✅, tasks: prev=10.0, new=20.6, diff=10.6 more (106.0%)
     q39: prev= 274 ms, new= 334 ms, diff=1.22 slower ✖, tasks: prev=0.0, new=9.2, diff=9.2 more
     q40: prev= 315 ms, new= 273 ms, diff=1.15 faster ✔, tasks: prev=2.0, new=7.0, diff=5.0 more (250.0%)
     q41: prev=  74 ms, new=  93 ms, diff=1.26 slower ✖, tasks: prev=0.0, new=4.0, diff=4.0 more
     q42: prev= 197 ms, new= 129 ms, diff=1.53 faster ✅, tasks: prev=2.0, new=6.0, diff=4.0 more (200.0%)
     q43: prev= 161 ms, new= 174 ms, diff=1.08 slower ✖, tasks: prev=0.0, new=4.0, diff=4.0 more
     q44: prev= 201 ms, new= 201 ms, diff=1.00 slower ✖, tasks: prev=8.0, new=17.0, diff=9.0 more (112.5%)
     q45: prev= 274 ms, new= 210 ms, diff=1.30 faster ✔, tasks: prev=0.0, new=6.0, diff=6.0 more
     q46: prev= 405 ms, new= 402 ms, diff=1.01 faster ✔, tasks: prev=0.0, new=8.0, diff=8.0 more
     q47: prev= 872 ms, new= 529 ms, diff=1.65 faster ✅, tasks: prev=0.0, new=15.4, diff=15.4 more
     q48: prev= 434 ms, new= 456 ms, diff=1.05 slower ✖, tasks: prev=0.0, new=4.0, diff=4.0 more
     q49: prev= 590 ms, new= 231 ms, diff=2.55 faster ✅, tasks: prev=9.0, new=26.0, diff=17.0 more (188.9%)
     q50: prev= 380 ms, new= 304 ms, diff=1.25 faster ✔, tasks: prev=2.0, new=7.0, diff=5.0 more (250.0%)
     q51: prev= 445 ms, new= 232 ms, diff=1.92 faster ✅, tasks: prev=14.0, new=30.4, diff=16.4 more (117.1%)
     q52: prev= 200 ms, new= 267 ms, diff=1.33 slower ✖, tasks: prev=2.0, new=6.0, diff=4.0 more (200.0%)
     q53: prev= 214 ms, new= 201 ms, diff=1.06 faster ✔, tasks: prev=0.0, new=5.0, diff=5.0 more
     q54: prev= 452 ms, new= 452 ms, diff=1.00 slower ✖, tasks: prev=2.0, new=15.0, diff=13.0 more (650.0%)
     q55: prev= 169 ms, new= 180 ms, diff=1.07 slower ✖, tasks: prev=0.0, new=4.0, diff=4.0 more
     q56: prev= 271 ms, new= 270 ms, diff=1.00 faster ✔, tasks: prev=0.0, new=15.0, diff=15.0 more
     q57: prev= 506 ms, new= 335 ms, diff=1.51 faster ✅, tasks: prev=0.0, new=13.2, diff=13.2 more
     q58: prev= 420 ms, new= 332 ms, diff=1.27 faster ✔, tasks: prev=0.0, new=18.0, diff=18.0 more
     q59: prev= 323 ms, new= 372 ms, diff=1.15 slower ✖, tasks: prev=0.0, new=8.0, diff=8.0 more
     q60: prev= 239 ms, new= 293 ms, diff=1.23 slower ✖, tasks: prev=0.0, new=15.0, diff=15.0 more
     q61: prev= 721 ms, new= 805 ms, diff=1.12 slower ✖, tasks: prev=0.0, new=13.0, diff=13.0 more
     q62: prev=1344 ms, new=1358 ms, diff=1.01 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
     q63: prev= 183 ms, new= 195 ms, diff=1.07 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
     q64: prev=1293 ms, new=1150 ms, diff=1.12 faster ✔, tasks: prev=18.0, new=58.6, diff=40.6 more (225.6%)
     q65: prev= 378 ms, new= 288 ms, diff=1.31 faster ✔, tasks: prev=0.0, new=8.0, diff=8.0 more
     q66: prev= 391 ms, new= 432 ms, diff=1.10 slower ✖, tasks: prev=0.0, new=9.0, diff=9.0 more
     q67: prev= 707 ms, new= 633 ms, diff=1.12 faster ✔, tasks: prev=2.0, new=19.2, diff=17.2 more (860.0%)
     q68: prev= 403 ms, new= 424 ms, diff=1.05 slower ✖, tasks: prev=0.0, new=8.0, diff=8.0 more
     q69: prev= 608 ms, new= 276 ms, diff=2.20 faster ✅, tasks: prev=0.0, new=13.0, diff=13.0 more
     q70: prev= 431 ms, new= 477 ms, diff=1.11 slower ✖, tasks: prev=0.0, new=7.0, diff=7.0 more
     q71: prev= 199 ms, new= 263 ms, diff=1.32 slower ✖, tasks: prev=0.0, new=7.0, diff=7.0 more
     q72: prev=13215 ms, new=12849 ms, diff=1.03 faster ✔, tasks: prev=2.0, new=17.0, diff=15.0 more (750.0%)
     q73: prev= 226 ms, new= 250 ms, diff=1.11 slower ✖, tasks: prev=0.0, new=6.0, diff=6.0 more
     q74: prev= 653 ms, new= 591 ms, diff=1.10 faster ✔, tasks: prev=0.0, new=12.0, diff=12.0 more
     q75: prev=1043 ms, new= 473 ms, diff=2.21 faster ✅, tasks: prev=10.0, new=46.2, diff=36.2 more (362.0%)
     q76: prev= 219 ms, new= 210 ms, diff=1.04 faster ✔, tasks: prev=5.0, new=10.0, diff=5.0 more (100.0%)
     q77: prev= 474 ms, new= 251 ms, diff=1.89 faster ✅, tasks: prev=5.0, new=20.0, diff=15.0 more (300.0%)
     q78: prev= 866 ms, new= 326 ms, diff=2.66 faster ✅, tasks: prev=19.0, new=48.0, diff=29.0 more (152.6%)
     q79: prev= 289 ms, new= 368 ms, diff=1.27 slower ✖, tasks: prev=0.0, new=6.0, diff=6.0 more
     q80: prev= 579 ms, new= 365 ms, diff=1.59 faster ✅, tasks: prev=9.0, new=27.0, diff=18.0 more (200.0%)
     q81: prev= 270 ms, new= 284 ms, diff=1.05 slower ✖, tasks: prev=0.0, new=11.0, diff=11.0 more
     q82: prev= 333 ms, new= 341 ms, diff=1.02 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
     q83: prev= 273 ms, new= 239 ms, diff=1.14 faster ✔, tasks: prev=0.0, new=16.4, diff=16.4 more
     q84: prev= 303 ms, new= 305 ms, diff=1.01 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
     q85: prev= 506 ms, new= 436 ms, diff=1.16 faster ✔, tasks: prev=0.0, new=14.0, diff=14.0 more
     q86: prev= 136 ms, new= 154 ms, diff=1.13 slower ✖, tasks: prev=0.0, new=4.0, diff=4.0 more
     q87: prev= 332 ms, new= 204 ms, diff=1.63 faster ✅, tasks: prev=10.0, new=20.6, diff=10.6 more (106.0%)
     q88: prev= 422 ms, new= 450 ms, diff=1.07 slower ✖, tasks: prev=0.0, new=32.0, diff=32.0 more
     q89: prev= 208 ms, new= 197 ms, diff=1.06 faster ✔, tasks: prev=0.0, new=6.0, diff=6.0 more
     q90: prev= 164 ms, new= 177 ms, diff=1.08 slower ✖, tasks: prev=0.0, new=8.0, diff=8.0 more
     q91: prev= 392 ms, new= 407 ms, diff=1.04 slower ✖, tasks: prev=0.0, new=7.0, diff=7.0 more
     q92: prev= 238 ms, new= 154 ms, diff=1.55 faster ✅, tasks: prev=0.0, new=6.0, diff=6.0 more
     q93: prev= 262 ms, new= 197 ms, diff=1.33 faster ✔, tasks: prev=2.0, new=7.0, diff=5.0 more (250.0%)
     q94: prev= 293 ms, new= 219 ms, diff=1.34 faster ✔, tasks: prev=0.0, new=9.8, diff=9.8 more
     q95: prev= 521 ms, new= 277 ms, diff=1.88 faster ✅, tasks: prev=0.0, new=27.2, diff=27.2 more
     q96: prev= 181 ms, new= 196 ms, diff=1.08 slower ✖, tasks: prev=0.0, new=4.0, diff=4.0 more
     q97: prev= 271 ms, new= 190 ms, diff=1.43 faster ✔, tasks: prev=0.0, new=7.0, diff=7.0 more
     q98: prev= 173 ms, new= 203 ms, diff=1.17 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
     q99: prev=14880 ms, new=15071 ms, diff=1.01 slower ✖, tasks: prev=0.0, new=5.0, diff=5.0 more
Verification and run details

Job 39 captured both immutable revisions when the request was queued. The bot fetched and checked out each full commit SHA in detached HEAD, then built and deployed the datafusion-distributed-benchmarks --bin worker target from that checkout.

Identity Base PR head
Source commit 81b887eb012980ad32a9de9da17b8109aedb49ee bc15154d43cf715db887a93aa942a63277416f0a
Phase Base PR head
Build and deployment 2m 22s 1m 29s
All benchmarks 7m 51s 7m 8s
Benchmark tpcds/sf1 7m 51s 7m 8s

Workload: tpcds/sf1 · all queries · 1 warmup + 5 measured iterations per query

Capacity: 12 c5n.2xlarge nodes for both revisions

Other timings: Queue 1s · Dataset validation 0s · Total 18m 58s

@gabotechs gabotechs closed this Aug 20, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants