You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@doris.apache.org by "kevinw66 (via GitHub)" <gi...@apache.org> on 2025/12/02 07:06:36 UTC
[PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
kevinw66 opened a new pull request, #58615:
URL: https://github.com/apache/doris/pull/58615
### What problem does this PR solve?
Issue Number: close #xxx
Related PR: #xxx
Problem Summary:
### Release note
None
### Check List (For Author)
- Test <!-- At least one of them must be included. -->
- [ ] Regression test
- [ ] Unit Test
- [ ] Manual test (add detailed scripts or steps below)
- [ ] No need to test or manual test. Explain why:
- [ ] This is a refactor/code format and no logic has been changed.
- [ ] Previous test can cover this change.
- [ ] No code files have been changed.
- [ ] Other reason <!-- Add your reason? -->
- Behavior changed:
- [ ] No.
- [ ] Yes. <!-- Explain the behavior change -->
- Does this need documentation?
- [ ] No.
- [ ] Yes. <!-- Add document PR link here. eg: https://github.com/apache/doris-website/pull/1214 -->
### Check List (For Reviewer who merge this PR)
- [ ] Confirm the release note
- [ ] Confirm test cases
- [ ] Confirm document
- [ ] Add branch pick label <!-- Add branch pick label that this PR should merge into -->
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3609930801
PR approved by at least one committer and no changes requested.
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3602377883
<details>
<summary>TPC-DS: <b>Total hot run time: 182393 ms</b></summary>
```
machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
scripts: https://github.com/apache/doris/tree/master/tools/tpcds-tools
TPC-DS sf100 test result on commit a8f3f968612c80a0a46d04d7dd6f338a2d87edd7, data reload: false
query1 1053 426 395 395
query2 6599 1190 1172 1172
query3 6759 227 220 220
query4 24932 23364 22993 22993
query5 5133 635 504 504
query6 352 266 236 236
query7 4658 532 312 312
query8 317 250 246 246
query9 8731 2673 2644 2644
query10 570 375 308 308
query11 15379 15117 14874 14874
query12 197 125 118 118
query13 1703 585 437 437
query14 9888 6020 6123 6020
query15 217 209 193 193
query16 7695 690 506 506
query17 1228 809 657 657
query18 2050 446 349 349
query19 226 218 194 194
query20 128 126 129 126
query21 238 153 129 129
query22 4012 4048 3872 3872
query23 32753 32071 31897 31897
query24 8478 2443 2425 2425
query25 632 572 499 499
query26 1263 289 171 171
query27 2718 506 346 346
query28 4345 2185 2187 2185
query29 839 647 525 525
query30 323 244 225 225
query31 856 722 619 619
query32 87 75 78 75
query33 617 438 347 347
query34 834 883 596 596
query35 775 812 747 747
query36 927 930 843 843
query37 123 114 91 91
query38 3937 3899 3787 3787
query39 1476 1411 1399 1399
query40 226 133 119 119
query41 65 62 65 62
query42 131 121 114 114
query43 437 467 422 422
query44 1326 765 745 745
query45 199 189 181 181
query46 898 1044 644 644
query47 1698 1733 1647 1647
query48 392 438 330 330
query49 805 499 419 419
query50 686 688 428 428
query51 3877 3977 3823 3823
query52 117 117 113 113
query53 245 260 193 193
query54 318 294 286 286
query55 97 94 93 93
query56 354 337 352 337
query57 1138 1153 1061 1061
query58 300 305 284 284
query59 2259 2464 2318 2318
query60 359 355 347 347
query61 160 163 165 163
query62 759 729 656 656
query63 235 195 197 195
query64 4501 1229 905 905
query65 4040 3929 3973 3929
query66 1096 439 338 338
query67 15058 14916 14804 14804
query68 8437 979 625 625
query69 516 348 318 318
query70 1116 1043 1024 1024
query71 474 362 312 312
query72 5642 5037 4833 4833
query73 713 587 347 347
query74 8660 8787 8708 8708
query75 3587 3047 2606 2606
query76 3841 1164 746 746
query77 819 404 336 336
query78 9383 9558 8926 8926
query79 1844 869 603 603
query80 702 585 509 509
query81 485 272 235 235
query82 242 141 114 114
query83 295 271 253 253
query84 307 123 107 107
query85 892 518 476 476
query86 342 320 277 277
query87 4071 4115 3964 3964
query88 2999 2360 2328 2328
query89 393 332 298 298
query90 2065 231 229 229
query91 173 169 146 146
query92 88 73 65 65
query93 1229 1045 665 665
query94 726 460 342 342
query95 508 417 409 409
query96 550 553 286 286
query97 2610 2666 2582 2582
query98 228 220 209 209
query99 1379 1385 1277 1277
Total cold run time: 270140 ms
Total hot run time: 182393 ms
```
</details>
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "hello-stephen (via GitHub)" <gi...@apache.org>.
hello-stephen commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3604971225
# BE Regression && UT Coverage Report
Increment line coverage `100.00% (15/15)` :tada:
[Increment coverage report](http://coverage.selectdb-in.cc/coverage/58615_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_merge/increment_report/index.html)
[Complete coverage report](http://coverage.selectdb-in.cc/coverage/58615_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_merge/report/index.html)
| Category | Coverage |
|-------------------|--------------------|
| Function Coverage | 72.32% (24823/34326) |
| Line Coverage | 58.96% (260194/441313) |
| Region Coverage | 53.82% (216073/401499) |
| Branch Coverage | 55.39% (92493/166979) |
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "Thearas (via GitHub)" <gi...@apache.org>.
Thearas commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3600508374
Thank you for your contribution to Apache Doris.
Don't know what should be done next? See [How to process your PR](https://cwiki.apache.org/confluence/display/DORIS/How+to+process+your+PR).
Please clearly describe your PR:
1. What problem was fixed (it's best to include specific error reporting information). How it was fixed.
2. Which behaviors were modified. What was the previous behavior, what is it now, why was it modified, and what possible impacts might there be.
3. What features were added. Why was this function added?
4. Which code was refactored and why was this part of the code refactored?
5. Which functions were optimized and what is the difference before and after the optimization?
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3609930866
PR approved by anyone and no changes requested.
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "kevinw66 (via GitHub)" <gi...@apache.org>.
kevinw66 commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3602057317
run buildall
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3603035844
<details>
<summary>ClickBench: <b>Total hot run time: 27.04 s</b></summary>
```
machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
scripts: https://github.com/apache/doris/tree/master/tools/clickbench-tools
ClickBench test result on commit 1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113, data reload: false
query1 0.06 0.05 0.05
query2 0.10 0.05 0.05
query3 0.26 0.08 0.09
query4 1.61 0.12 0.11
query5 0.28 0.26 0.26
query6 1.18 0.64 0.62
query7 0.04 0.03 0.02
query8 0.06 0.04 0.04
query9 0.55 0.50 0.50
query10 0.56 0.55 0.55
query11 0.16 0.11 0.12
query12 0.15 0.13 0.12
query13 0.63 0.60 0.59
query14 0.98 0.98 0.99
query15 0.82 0.81 0.80
query16 0.41 0.40 0.37
query17 1.03 1.09 1.03
query18 0.23 0.22 0.21
query19 1.88 1.83 1.79
query20 0.02 0.01 0.01
query21 15.45 0.27 0.14
query22 4.94 0.05 0.05
query23 15.93 0.30 0.11
query24 1.19 0.60 0.19
query25 0.09 0.06 0.06
query26 0.15 0.13 0.14
query27 0.08 0.04 0.05
query28 4.42 1.21 1.02
query29 12.56 3.97 3.14
query30 0.27 0.14 0.12
query31 2.82 0.61 0.39
query32 3.23 0.57 0.45
query33 3.06 3.06 3.07
query34 16.61 5.28 4.49
query35 4.55 4.56 4.52
query36 0.66 0.49 0.49
query37 0.10 0.06 0.07
query38 0.08 0.04 0.03
query39 0.05 0.03 0.03
query40 0.17 0.15 0.14
query41 0.09 0.03 0.03
query42 0.04 0.03 0.03
query43 0.04 0.04 0.03
Total cold run time: 97.59 s
Total hot run time: 27.04 s
```
</details>
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "hello-stephen (via GitHub)" <gi...@apache.org>.
hello-stephen commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3605054512
# BE Regression && UT Coverage Report
Increment line coverage `100.00% (15/15)` :tada:
[Increment coverage report](http://coverage.selectdb-in.cc/coverage/58615_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_merge/increment_report/index.html)
[Complete coverage report](http://coverage.selectdb-in.cc/coverage/58615_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_merge/report/index.html)
| Category | Coverage |
|-------------------|--------------------|
| Function Coverage | 72.32% (24823/34326) |
| Line Coverage | 58.96% (260194/441313) |
| Region Coverage | 53.82% (216073/401499) |
| Branch Coverage | 55.39% (92493/166979) |
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3600738345
<details>
<summary>TPC-DS: <b>Total hot run time: 181890 ms</b></summary>
```
machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
scripts: https://github.com/apache/doris/tree/master/tools/tpcds-tools
TPC-DS sf100 test result on commit 7e407061893364938b5613c5f4748565fb7fc27a, data reload: false
query1 1075 407 398 398
query2 6589 1206 1188 1188
query3 6740 236 230 230
query4 25417 23152 22980 22980
query5 5036 658 528 528
query6 330 242 230 230
query7 4684 515 313 313
query8 312 246 246 246
query9 8737 2628 2630 2628
query10 547 366 345 345
query11 15114 15251 14519 14519
query12 196 126 114 114
query13 1683 561 427 427
query14 9555 5882 5842 5842
query15 205 204 186 186
query16 7690 724 550 550
query17 1174 758 610 610
query18 2049 423 366 366
query19 207 202 175 175
query20 128 132 120 120
query21 216 135 121 121
query22 3893 3948 3782 3782
query23 32948 31921 31719 31719
query24 8556 2397 2406 2397
query25 614 535 478 478
query26 1246 276 167 167
query27 2745 524 347 347
query28 4370 2175 2157 2157
query29 836 635 506 506
query30 303 247 220 220
query31 801 715 647 647
query32 86 74 71 71
query33 604 394 332 332
query34 836 886 543 543
query35 784 843 735 735
query36 925 936 833 833
query37 130 116 92 92
query38 3935 3848 3738 3738
query39 1502 1423 1440 1423
query40 222 135 126 126
query41 73 76 62 62
query42 131 122 118 118
query43 457 452 428 428
query44 1335 751 757 751
query45 202 197 187 187
query46 895 1012 649 649
query47 1706 1721 1647 1647
query48 405 422 322 322
query49 817 505 419 419
query50 713 707 411 411
query51 3876 3954 4002 3954
query52 118 114 109 109
query53 249 263 195 195
query54 324 314 310 310
query55 100 107 96 96
query56 350 354 349 349
query57 1172 1162 1090 1090
query58 303 292 290 290
query59 2342 2459 2410 2410
query60 384 374 338 338
query61 193 194 190 190
query62 800 736 655 655
query63 233 199 209 199
query64 4667 1344 1063 1063
query65 4094 3984 4025 3984
query66 1128 476 364 364
query67 15326 14993 14863 14863
query68 7124 986 634 634
query69 533 354 322 322
query70 1107 1037 1053 1037
query71 427 351 319 319
query72 6075 4889 4862 4862
query73 691 589 347 347
query74 8499 8828 8769 8769
query75 3032 3021 2474 2474
query76 3316 1141 751 751
query77 507 416 315 315
query78 9455 9709 8852 8852
query79 1934 882 603 603
query80 723 608 510 510
query81 528 273 249 249
query82 432 140 115 115
query83 276 268 251 251
query84 262 126 97 97
query85 907 504 465 465
query86 387 308 275 275
query87 4064 4039 3961 3961
query88 3107 2341 2316 2316
query89 390 333 296 296
query90 1820 236 229 229
query91 176 169 145 145
query92 76 72 70 70
query93 1318 1022 661 661
query94 689 459 352 352
query95 523 428 414 414
query96 553 546 287 287
query97 2608 2703 2555 2555
query98 243 220 208 208
query99 1316 1390 1267 1267
Total cold run time: 268511 ms
Total hot run time: 181890 ms
```
</details>
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3603011454
<details>
<summary>TPC-DS: <b>Total hot run time: 181400 ms</b></summary>
```
machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
scripts: https://github.com/apache/doris/tree/master/tools/tpcds-tools
TPC-DS sf100 test result on commit 1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113, data reload: false
query1 1040 406 398 398
query2 6600 1214 1197 1197
query3 6744 224 226 224
query4 25313 23176 22917 22917
query5 4842 639 496 496
query6 342 248 214 214
query7 4663 525 307 307
query8 319 261 259 259
query9 8708 2631 2651 2631
query10 527 358 312 312
query11 15524 14793 14541 14541
query12 179 120 116 116
query13 1719 605 443 443
query14 9447 6227 6097 6097
query15 212 204 178 178
query16 7557 692 531 531
query17 1217 785 651 651
query18 2032 440 337 337
query19 229 212 190 190
query20 128 128 122 122
query21 221 133 114 114
query22 3871 3870 3826 3826
query23 32855 31931 31861 31861
query24 8493 2409 2423 2409
query25 621 505 457 457
query26 1253 270 171 171
query27 2762 510 343 343
query28 4311 2157 2139 2139
query29 809 622 480 480
query30 307 241 205 205
query31 805 705 627 627
query32 87 75 71 71
query33 609 383 330 330
query34 832 887 537 537
query35 769 841 752 752
query36 884 943 848 848
query37 131 112 94 94
query38 3879 3857 3753 3753
query39 1486 1396 1382 1382
query40 221 131 121 121
query41 64 61 61 61
query42 129 114 114 114
query43 443 445 426 426
query44 1319 741 754 741
query45 197 193 182 182
query46 899 1027 647 647
query47 1653 1704 1627 1627
query48 405 430 337 337
query49 776 515 415 415
query50 678 696 419 419
query51 3825 3933 3870 3870
query52 112 115 105 105
query53 236 262 195 195
query54 316 299 282 282
query55 97 96 94 94
query56 348 329 304 304
query57 1138 1162 1089 1089
query58 290 277 267 267
query59 2329 2370 2260 2260
query60 353 358 351 351
query61 163 158 161 158
query62 770 706 651 651
query63 233 191 197 191
query64 4598 1186 897 897
query65 4042 3980 3987 3980
query66 1152 428 328 328
query67 14921 14796 14878 14796
query68 5552 1013 629 629
query69 522 342 308 308
query70 1105 1044 1036 1036
query71 429 338 324 324
query72 5903 4820 4767 4767
query73 671 587 357 357
query74 8742 8797 8525 8525
query75 3051 3103 2573 2573
query76 3333 1146 758 758
query77 519 409 329 329
query78 9543 9753 8888 8888
query79 1634 832 583 583
query80 1694 630 540 540
query81 559 287 253 253
query82 407 142 124 124
query83 381 272 271 271
query84 266 126 111 111
query85 1068 492 442 442
query86 392 306 285 285
query87 4011 4064 3986 3986
query88 3068 2320 2348 2320
query89 406 343 294 294
query90 1752 246 217 217
query91 165 177 150 150
query92 72 72 65 65
query93 1222 1015 673 673
query94 725 436 352 352
query95 510 411 407 407
query96 537 557 285 285
query97 2561 2691 2618 2618
query98 252 224 207 207
query99 1320 1388 1255 1255
Total cold run time: 266406 ms
Total hot run time: 181400 ms
```
</details>
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "kevinw66 (via GitHub)" <gi...@apache.org>.
kevinw66 commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3600514051
## Improvements
### BM_Bits_CountZeroNum
| Size | Origin (ns) | NEON (ns) | Improvement |
|------|-------------|-----------|-------------|
| 16 | 5.44 | 5.44 | 0.0% |
| 32 | 10.3 | 10.3 | 0.0% |
| 64 | 17.3 | 1.47 | +91.5% |
| 256 | 59.2 | 4.71 | +92.0% |
| 1024 | 227 | 18.4 | +91.9% |
### BM_Bits_CountZeroNumNullMap
| Size | Origin (ns) | NEON (ns) | Improvement |
|------|-------------|-----------|-------------|
| 16 | 3.40 | 3.39 | +0.3% |
| 32 | 7.29 | 7.28 | +0.1% |
| 64 | 10.4 | 2.65 | +74.4% |
| 256 | 29.9 | 8.94 | +70.2% |
| 1024 | 107 | 34.1 | +68.1% |
Summary: NEON matches origin at very small sizes (16–32B) and delivers large gains for 64B+ (68–92%).
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3600755651
<details>
<summary>ClickBench: <b>Total hot run time: 27.41 s</b></summary>
```
machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
scripts: https://github.com/apache/doris/tree/master/tools/clickbench-tools
ClickBench test result on commit 7e407061893364938b5613c5f4748565fb7fc27a, data reload: false
query1 0.05 0.05 0.05
query2 0.11 0.05 0.05
query3 0.26 0.08 0.08
query4 1.62 0.12 0.11
query5 0.28 0.27 0.24
query6 1.15 0.64 0.64
query7 0.03 0.03 0.03
query8 0.05 0.04 0.04
query9 0.56 0.50 0.50
query10 0.56 0.55 0.56
query11 0.15 0.12 0.11
query12 0.14 0.11 0.12
query13 0.62 0.61 0.60
query14 1.01 0.98 1.01
query15 0.81 0.81 0.80
query16 0.39 0.38 0.39
query17 1.06 1.08 1.05
query18 0.22 0.22 0.22
query19 1.90 1.86 1.82
query20 0.01 0.01 0.02
query21 15.45 0.27 0.14
query22 5.06 0.05 0.05
query23 15.95 0.28 0.11
query24 1.47 0.36 0.34
query25 0.08 0.07 0.06
query26 0.17 0.14 0.13
query27 0.10 0.05 0.08
query28 3.94 1.21 1.02
query29 12.57 3.90 3.17
query30 0.28 0.14 0.12
query31 2.82 0.62 0.40
query32 3.23 0.56 0.46
query33 2.97 3.11 3.03
query34 16.88 5.18 4.58
query35 4.54 4.52 4.57
query36 0.65 0.50 0.49
query37 0.11 0.07 0.06
query38 0.08 0.04 0.04
query39 0.05 0.03 0.03
query40 0.17 0.14 0.14
query41 0.10 0.03 0.03
query42 0.04 0.04 0.03
query43 0.04 0.04 0.04
Total cold run time: 97.73 s
Total hot run time: 27.41 s
```
</details>
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3600699368
<details>
<summary>TPC-H: <b>Total hot run time: 34467 ms</b></summary>
```
machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
scripts: https://github.com/apache/doris/tree/master/tools/tpch-tools
Tpch sf100 test result on commit 7e407061893364938b5613c5f4748565fb7fc27a, data reload: false
------ Round 1 ----------------------------------
q1 17658 5032 4905 4905
q2 2062 331 209 209
q3 10242 1327 735 735
q4 10219 865 323 323
q5 7545 2461 2180 2180
q6 198 171 136 136
q7 966 796 647 647
q8 9358 1448 1173 1173
q9 7165 5371 5318 5318
q10 6798 2235 1788 1788
q11 530 312 282 282
q12 338 380 220 220
q13 17767 3717 3021 3021
q14 242 240 217 217
q15 577 532 519 519
q16 895 877 809 809
q17 670 764 566 566
q18 7749 7244 7071 7071
q19 1101 979 616 616
q20 357 364 230 230
q21 4070 3716 2511 2511
q22 1045 991 993 991
Total cold run time: 107552 ms
Total hot run time: 34467 ms
----- Round 2, with runtime_filter_mode=off -----
q1 4924 4917 4902 4902
q2 311 370 308 308
q3 2154 2700 2249 2249
q4 1340 1767 1290 1290
q5 4258 4624 4616 4616
q6 216 170 128 128
q7 2058 1961 1855 1855
q8 2680 2537 2530 2530
q9 7720 7457 7515 7457
q10 3071 3265 2872 2872
q11 601 529 499 499
q12 742 771 659 659
q13 3596 4124 3225 3225
q14 308 303 276 276
q15 552 510 499 499
q16 893 946 895 895
q17 1294 1572 1431 1431
q18 7974 7861 7526 7526
q19 903 882 908 882
q20 2076 2038 1956 1956
q21 4936 4259 4246 4246
q22 1123 1005 983 983
Total cold run time: 53730 ms
Total hot run time: 51284 ms
```
</details>
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3602325958
<details>
<summary>TPC-H: <b>Total hot run time: 35768 ms</b></summary>
```
machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
scripts: https://github.com/apache/doris/tree/master/tools/tpch-tools
Tpch sf100 test result on commit a8f3f968612c80a0a46d04d7dd6f338a2d87edd7, data reload: false
------ Round 1 ----------------------------------
q1 17629 5251 4990 4990
q2 2038 329 216 216
q3 10261 1347 749 749
q4 10228 883 317 317
q5 7507 2448 2159 2159
q6 195 178 145 145
q7 962 781 645 645
q8 9354 1495 1113 1113
q9 7077 5312 5324 5312
q10 6828 2205 1764 1764
q11 517 309 297 297
q12 344 372 227 227
q13 17800 3698 3031 3031
q14 234 234 216 216
q15 593 523 516 516
q16 894 882 821 821
q17 691 770 596 596
q18 7418 7238 7563 7238
q19 1116 1034 694 694
q20 432 398 247 247
q21 4326 3968 3439 3439
q22 1156 1036 1070 1036
Total cold run time: 107600 ms
Total hot run time: 35768 ms
----- Round 2, with runtime_filter_mode=off -----
q1 5316 5232 5288 5232
q2 356 395 337 337
q3 2457 2869 2466 2466
q4 1422 1915 1359 1359
q5 4535 4505 4449 4449
q6 212 168 133 133
q7 2065 2003 1771 1771
q8 2681 2480 2591 2480
q9 7364 7630 7024 7024
q10 2907 3132 2673 2673
q11 574 511 491 491
q12 650 732 558 558
q13 3291 3695 3113 3113
q14 286 293 265 265
q15 544 494 508 494
q16 897 888 856 856
q17 1122 1434 1394 1394
q18 7134 7106 7110 7106
q19 846 790 844 790
q20 1892 1955 1902 1902
q21 4696 4308 4206 4206
q22 1102 1017 977 977
Total cold run time: 52349 ms
Total hot run time: 50076 ms
```
</details>
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "hello-stephen (via GitHub)" <gi...@apache.org>.
hello-stephen commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3605154771
# BE Regression && UT Coverage Report
Increment line coverage `100.00% (15/15)` :tada:
[Increment coverage report](http://coverage.selectdb-in.cc/coverage/58615_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_merge/increment_report/index.html)
[Complete coverage report](http://coverage.selectdb-in.cc/coverage/58615_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_merge/report/index.html)
| Category | Coverage |
|-------------------|--------------------|
| Function Coverage | 72.31% (24820/34326) |
| Line Coverage | 58.94% (260096/441313) |
| Region Coverage | 53.78% (215945/401499) |
| Branch Coverage | 55.36% (92437/166979) |
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "kevinw66 (via GitHub)" <gi...@apache.org>.
kevinw66 commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3602767046
run buildall
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "zclllyybb (via GitHub)" <gi...@apache.org>.
zclllyybb merged PR #58615:
URL: https://github.com/apache/doris/pull/58615
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3605970804
<details>
<summary>TPC-DS: <b>Total hot run time: 183409 ms</b></summary>
```
machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
scripts: https://github.com/apache/doris/tree/master/tools/tpcds-tools
TPC-DS sf100 test result on commit 1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113, data reload: false
query1 1030 407 405 405
query2 6627 1199 1173 1173
query3 6744 237 229 229
query4 25769 23277 22806 22806
query5 4848 670 507 507
query6 364 253 236 236
query7 4665 529 323 323
query8 318 262 247 247
query9 8730 2655 2661 2655
query10 543 381 316 316
query11 15525 14965 15145 14965
query12 193 123 118 118
query13 1706 598 455 455
query14 10100 6353 6237 6237
query15 226 204 193 193
query16 7694 721 556 556
query17 1633 824 659 659
query18 2058 460 354 354
query19 225 210 190 190
query20 136 136 131 131
query21 227 150 122 122
query22 3948 4041 4020 4020
query23 33128 32206 31897 31897
query24 8377 2464 2492 2464
query25 646 519 472 472
query26 1245 276 173 173
query27 2710 502 355 355
query28 4334 2230 2214 2214
query29 788 624 491 491
query30 324 247 218 218
query31 827 717 643 643
query32 82 77 75 75
query33 606 391 346 346
query34 858 904 551 551
query35 789 848 759 759
query36 896 924 853 853
query37 132 118 95 95
query38 3975 3935 3870 3870
query39 1487 1409 1433 1409
query40 238 139 126 126
query41 67 64 63 63
query42 135 119 114 114
query43 469 468 435 435
query44 1368 801 777 777
query45 209 202 185 185
query46 906 1024 661 661
query47 1691 1706 1641 1641
query48 428 438 341 341
query49 816 529 419 419
query50 725 723 430 430
query51 3962 4078 3966 3966
query52 129 118 110 110
query53 264 280 203 203
query54 321 309 290 290
query55 102 98 100 98
query56 387 335 333 333
query57 1131 1145 1085 1085
query58 303 285 287 285
query59 2343 2438 2231 2231
query60 358 364 361 361
query61 170 168 167 167
query62 794 721 660 660
query63 237 198 201 198
query64 4482 1269 957 957
query65 4084 4056 3972 3972
query66 1097 504 363 363
query67 15194 15066 14710 14710
query68 6753 1021 669 669
query69 532 361 320 320
query70 1127 1055 1042 1042
query71 452 367 331 331
query72 5943 5040 5033 5033
query73 680 580 363 363
query74 8892 8772 8569 8569
query75 3059 3106 2550 2550
query76 3346 1154 783 783
query77 517 461 326 326
query78 9502 9844 8936 8936
query79 2162 848 600 600
query80 1661 596 520 520
query81 580 272 242 242
query82 456 148 123 123
query83 372 277 273 273
query84 271 122 105 105
query85 934 522 449 449
query86 469 308 280 280
query87 4103 4143 4024 4024
query88 3154 2331 2318 2318
query89 392 341 310 310
query90 1915 234 232 232
query91 177 176 145 145
query92 81 72 69 69
query93 1637 1064 688 688
query94 780 443 358 358
query95 530 434 408 408
query96 544 574 291 291
query97 2628 2688 2587 2587
query98 249 234 215 215
query99 1325 1404 1287 1287
Total cold run time: 272146 ms
Total hot run time: 183409 ms
```
</details>
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3605990061
<details>
<summary>ClickBench: <b>Total hot run time: 27.72 s</b></summary>
```
machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
scripts: https://github.com/apache/doris/tree/master/tools/clickbench-tools
ClickBench test result on commit 1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113, data reload: false
query1 0.06 0.04 0.05
query2 0.10 0.05 0.05
query3 0.26 0.09 0.10
query4 1.62 0.12 0.11
query5 0.27 0.26 0.27
query6 1.16 0.65 0.64
query7 0.03 0.03 0.03
query8 0.06 0.04 0.04
query9 0.58 0.51 0.51
query10 0.56 0.55 0.56
query11 0.17 0.11 0.11
query12 0.15 0.12 0.11
query13 0.62 0.60 0.60
query14 1.00 0.98 1.00
query15 0.83 0.80 0.81
query16 0.41 0.39 0.40
query17 1.02 1.03 1.01
query18 0.25 0.22 0.25
query19 1.94 1.90 1.79
query20 0.02 0.01 0.02
query21 15.45 0.29 0.14
query22 4.65 0.06 0.05
query23 16.04 0.30 0.10
query24 1.29 0.51 0.25
query25 0.08 0.08 0.06
query26 0.15 0.14 0.14
query27 0.06 0.06 0.05
query28 3.24 1.24 1.03
query29 12.58 4.08 3.38
query30 0.27 0.14 0.13
query31 2.82 0.63 0.41
query32 3.24 0.56 0.47
query33 2.99 3.12 3.10
query34 17.17 5.29 4.60
query35 4.62 4.58 4.63
query36 0.70 0.53 0.52
query37 0.11 0.07 0.06
query38 0.07 0.04 0.05
query39 0.05 0.03 0.03
query40 0.18 0.15 0.14
query41 0.09 0.04 0.04
query42 0.05 0.03 0.03
query43 0.04 0.04 0.03
Total cold run time: 97.05 s
Total hot run time: 27.72 s
```
</details>
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "hello-stephen (via GitHub)" <gi...@apache.org>.
hello-stephen commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3605096461
# BE Regression && UT Coverage Report
Increment line coverage `100.00% (15/15)` :tada:
[Increment coverage report](http://coverage.selectdb-in.cc/coverage/58615_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_merge/increment_report/index.html)
[Complete coverage report](http://coverage.selectdb-in.cc/coverage/58615_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_merge/report/index.html)
| Category | Coverage |
|-------------------|--------------------|
| Function Coverage | 72.32% (24823/34326) |
| Line Coverage | 58.95% (260174/441313) |
| Region Coverage | 53.80% (216003/401499) |
| Branch Coverage | 55.38% (92477/166979) |
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "kevinw66 (via GitHub)" <gi...@apache.org>.
kevinw66 commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3600510777
Benchmark results on Huawei kC2 instance:
Original:
```
Running ./output/be/lib/benchmark_test
Run on (16 X 200 MHz CPU s)
CPU Caches:
L1 Data 64 KiB (x16)
L1 Instruction 64 KiB (x16)
L2 Unified 512 KiB (x16)
L3 Unified 32768 KiB (x1)
Load Average: 1.55, 8.69, 12.21
------------------------------------------------------------------------------------------------------------
Benchmark Time CPU Iterations UserCounters...
------------------------------------------------------------------------------------------------------------
BM_Bits_CountZeroNum/16/repeats:5_mean 5.44 ns 5.44 ns 5 bytes_per_second=2.74127G/s
BM_Bits_CountZeroNum/16/repeats:5_median 5.44 ns 5.44 ns 5 bytes_per_second=2.7413G/s
BM_Bits_CountZeroNum/16/repeats:5_stddev 0.000 ns 0.000 ns 5 bytes_per_second=213.303k/s
BM_Bits_CountZeroNum/16/repeats:5_cv 0.01 % 0.01 % 5 bytes_per_second=0.01%
BM_Bits_CountZeroNum/32/repeats:5_mean 10.3 ns 10.3 ns 5 bytes_per_second=2.8914G/s
BM_Bits_CountZeroNum/32/repeats:5_median 10.3 ns 10.3 ns 5 bytes_per_second=2.89148G/s
BM_Bits_CountZeroNum/32/repeats:5_stddev 0.001 ns 0.001 ns 5 bytes_per_second=236.223k/s
BM_Bits_CountZeroNum/32/repeats:5_cv 0.01 % 0.01 % 5 bytes_per_second=0.01%
BM_Bits_CountZeroNum/64/repeats:5_mean 17.3 ns 17.3 ns 5 bytes_per_second=3.4399G/s
BM_Bits_CountZeroNum/64/repeats:5_median 17.3 ns 17.3 ns 5 bytes_per_second=3.44002G/s
BM_Bits_CountZeroNum/64/repeats:5_stddev 0.002 ns 0.002 ns 5 bytes_per_second=420.005k/s
BM_Bits_CountZeroNum/64/repeats:5_cv 0.01 % 0.01 % 5 bytes_per_second=0.01%
BM_Bits_CountZeroNum/256/repeats:5_mean 59.2 ns 59.2 ns 5 bytes_per_second=4.03016G/s
BM_Bits_CountZeroNum/256/repeats:5_median 59.2 ns 59.2 ns 5 bytes_per_second=4.03013G/s
BM_Bits_CountZeroNum/256/repeats:5_stddev 0.003 ns 0.003 ns 5 bytes_per_second=214.616k/s
BM_Bits_CountZeroNum/256/repeats:5_cv 0.01 % 0.01 % 5 bytes_per_second=0.01%
BM_Bits_CountZeroNum/1024/repeats:5_mean 227 ns 226 ns 5 bytes_per_second=4.2107G/s
BM_Bits_CountZeroNum/1024/repeats:5_median 227 ns 226 ns 5 bytes_per_second=4.21062G/s
BM_Bits_CountZeroNum/1024/repeats:5_stddev 0.014 ns 0.014 ns 5 bytes_per_second=264.002k/s
BM_Bits_CountZeroNum/1024/repeats:5_cv 0.01 % 0.01 % 5 bytes_per_second=0.01%
BM_Bits_CountZeroNumNullMap/16/repeats:5_mean 3.40 ns 3.40 ns 5 bytes_per_second=4.38851G/s
BM_Bits_CountZeroNumNullMap/16/repeats:5_median 3.39 ns 3.39 ns 5 bytes_per_second=4.39511G/s
BM_Bits_CountZeroNumNullMap/16/repeats:5_stddev 0.011 ns 0.011 ns 5 bytes_per_second=15.1472M/s
BM_Bits_CountZeroNumNullMap/16/repeats:5_cv 0.34 % 0.34 % 5 bytes_per_second=0.34%
BM_Bits_CountZeroNumNullMap/32/repeats:5_mean 7.29 ns 7.29 ns 5 bytes_per_second=4.09028G/s
BM_Bits_CountZeroNumNullMap/32/repeats:5_median 7.29 ns 7.29 ns 5 bytes_per_second=4.08916G/s
BM_Bits_CountZeroNumNullMap/32/repeats:5_stddev 0.005 ns 0.005 ns 5 bytes_per_second=2.76443M/s
BM_Bits_CountZeroNumNullMap/32/repeats:5_cv 0.07 % 0.07 % 5 bytes_per_second=0.07%
BM_Bits_CountZeroNumNullMap/64/repeats:5_mean 10.4 ns 10.4 ns 5 bytes_per_second=5.7434G/s
BM_Bits_CountZeroNumNullMap/64/repeats:5_median 10.4 ns 10.4 ns 5 bytes_per_second=5.74335G/s
BM_Bits_CountZeroNumNullMap/64/repeats:5_stddev 0.001 ns 0.001 ns 5 bytes_per_second=668.068k/s
BM_Bits_CountZeroNumNullMap/64/repeats:5_cv 0.01 % 0.01 % 5 bytes_per_second=0.01%
BM_Bits_CountZeroNumNullMap/256/repeats:5_mean 29.9 ns 29.9 ns 5 bytes_per_second=7.98586G/s
BM_Bits_CountZeroNumNullMap/256/repeats:5_median 29.9 ns 29.9 ns 5 bytes_per_second=7.98601G/s
BM_Bits_CountZeroNumNullMap/256/repeats:5_stddev 0.006 ns 0.006 ns 5 bytes_per_second=1.59699M/s
BM_Bits_CountZeroNumNullMap/256/repeats:5_cv 0.02 % 0.02 % 5 bytes_per_second=0.02%
BM_Bits_CountZeroNumNullMap/1024/repeats:5_mean 107 ns 107 ns 5 bytes_per_second=8.89231G/s
BM_Bits_CountZeroNumNullMap/1024/repeats:5_median 107 ns 107 ns 5 bytes_per_second=8.8917G/s
BM_Bits_CountZeroNumNullMap/1024/repeats:5_stddev 0.016 ns 0.016 ns 5 bytes_per_second=1.34061M/s
BM_Bits_CountZeroNumNullMap/1024/repeats:5_cv 0.01 % 0.01 % 5 bytes_per_second=0.01%
```
Neon:
```
Running ./output/be/lib/benchmark_test
Run on (16 X 200 MHz CPU s)
CPU Caches:
L1 Data 64 KiB (x16)
L1 Instruction 64 KiB (x16)
L2 Unified 512 KiB (x16)
L3 Unified 32768 KiB (x1)
Load Average: 0.08, 0.98, 3.99
------------------------------------------------------------------------------------------------------------
Benchmark Time CPU Iterations UserCounters...
------------------------------------------------------------------------------------------------------------
BM_Bits_CountZeroNum/16/repeats:5_mean 5.44 ns 5.44 ns 5 bytes_per_second=2.74146G/s
BM_Bits_CountZeroNum/16/repeats:5_median 5.44 ns 5.44 ns 5 bytes_per_second=2.74143G/s
BM_Bits_CountZeroNum/16/repeats:5_stddev 0.001 ns 0.001 ns 5 bytes_per_second=337.724k/s
BM_Bits_CountZeroNum/16/repeats:5_cv 0.01 % 0.01 % 5 bytes_per_second=0.01%
BM_Bits_CountZeroNum/32/repeats:5_mean 10.3 ns 10.3 ns 5 bytes_per_second=2.89053G/s
BM_Bits_CountZeroNum/32/repeats:5_median 10.3 ns 10.3 ns 5 bytes_per_second=2.89058G/s
BM_Bits_CountZeroNum/32/repeats:5_stddev 0.000 ns 0.000 ns 5 bytes_per_second=107.791k/s
BM_Bits_CountZeroNum/32/repeats:5_cv 0.00 % 0.00 % 5 bytes_per_second=0.00%
BM_Bits_CountZeroNum/64/repeats:5_mean 1.47 ns 1.47 ns 5 bytes_per_second=40.5144G/s
BM_Bits_CountZeroNum/64/repeats:5_median 1.47 ns 1.47 ns 5 bytes_per_second=40.5145G/s
BM_Bits_CountZeroNum/64/repeats:5_stddev 0.000 ns 0.000 ns 5 bytes_per_second=3.31223M/s
BM_Bits_CountZeroNum/64/repeats:5_cv 0.01 % 0.01 % 5 bytes_per_second=0.01%
BM_Bits_CountZeroNum/256/repeats:5_mean 4.71 ns 4.71 ns 5 bytes_per_second=50.6679G/s
BM_Bits_CountZeroNum/256/repeats:5_median 4.65 ns 4.65 ns 5 bytes_per_second=51.2558G/s
BM_Bits_CountZeroNum/256/repeats:5_stddev 0.128 ns 0.128 ns 5 bytes_per_second=1.32569G/s
BM_Bits_CountZeroNum/256/repeats:5_cv 2.71 % 2.71 % 5 bytes_per_second=2.62%
BM_Bits_CountZeroNum/1024/repeats:5_mean 18.4 ns 18.4 ns 5 bytes_per_second=51.7001G/s
BM_Bits_CountZeroNum/1024/repeats:5_median 18.4 ns 18.4 ns 5 bytes_per_second=51.9428G/s
BM_Bits_CountZeroNum/1024/repeats:5_stddev 0.201 ns 0.201 ns 5 bytes_per_second=569.203M/s
BM_Bits_CountZeroNum/1024/repeats:5_cv 1.09 % 1.09 % 5 bytes_per_second=1.08%
BM_Bits_CountZeroNumNullMap/16/repeats:5_mean 3.39 ns 3.39 ns 5 bytes_per_second=4.3946G/s
BM_Bits_CountZeroNumNullMap/16/repeats:5_median 3.39 ns 3.39 ns 5 bytes_per_second=4.39535G/s
BM_Bits_CountZeroNumNullMap/16/repeats:5_stddev 0.002 ns 0.002 ns 5 bytes_per_second=2.13898M/s
BM_Bits_CountZeroNumNullMap/16/repeats:5_cv 0.05 % 0.05 % 5 bytes_per_second=0.05%
BM_Bits_CountZeroNumNullMap/32/repeats:5_mean 7.28 ns 7.28 ns 5 bytes_per_second=4.09179G/s
BM_Bits_CountZeroNumNullMap/32/repeats:5_median 7.29 ns 7.29 ns 5 bytes_per_second=4.0897G/s
BM_Bits_CountZeroNumNullMap/32/repeats:5_stddev 0.007 ns 0.007 ns 5 bytes_per_second=3.75246M/s
BM_Bits_CountZeroNumNullMap/32/repeats:5_cv 0.09 % 0.09 % 5 bytes_per_second=0.09%
BM_Bits_CountZeroNumNullMap/64/repeats:5_mean 2.65 ns 2.65 ns 5 bytes_per_second=22.46G/s
BM_Bits_CountZeroNumNullMap/64/repeats:5_median 2.65 ns 2.65 ns 5 bytes_per_second=22.4601G/s
BM_Bits_CountZeroNumNullMap/64/repeats:5_stddev 0.000 ns 0.000 ns 5 bytes_per_second=733.167k/s
BM_Bits_CountZeroNumNullMap/64/repeats:5_cv 0.00 % 0.00 % 5 bytes_per_second=0.00%
BM_Bits_CountZeroNumNullMap/256/repeats:5_mean 8.94 ns 8.94 ns 5 bytes_per_second=26.6831G/s
BM_Bits_CountZeroNumNullMap/256/repeats:5_median 8.89 ns 8.89 ns 5 bytes_per_second=26.8247G/s
BM_Bits_CountZeroNumNullMap/256/repeats:5_stddev 0.109 ns 0.109 ns 5 bytes_per_second=327.742M/s
BM_Bits_CountZeroNumNullMap/256/repeats:5_cv 1.22 % 1.22 % 5 bytes_per_second=1.20%
BM_Bits_CountZeroNumNullMap/1024/repeats:5_mean 34.1 ns 34.1 ns 5 bytes_per_second=27.9668G/s
BM_Bits_CountZeroNumNullMap/1024/repeats:5_median 33.9 ns 33.9 ns 5 bytes_per_second=28.1313G/s
BM_Bits_CountZeroNumNullMap/1024/repeats:5_stddev 0.462 ns 0.462 ns 5 bytes_per_second=380.851M/s
BM_Bits_CountZeroNumNullMap/1024/repeats:5_cv 1.35 % 1.35 % 5 bytes_per_second=1.33%
```
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "kevinw66 (via GitHub)" <gi...@apache.org>.
kevinw66 commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3600515458
run buildall
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3602400913
<details>
<summary>ClickBench: <b>Total hot run time: 27.21 s</b></summary>
```
machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
scripts: https://github.com/apache/doris/tree/master/tools/clickbench-tools
ClickBench test result on commit a8f3f968612c80a0a46d04d7dd6f338a2d87edd7, data reload: false
query1 0.05 0.05 0.05
query2 0.10 0.05 0.05
query3 0.25 0.09 0.09
query4 1.60 0.12 0.11
query5 0.27 0.28 0.27
query6 1.17 0.64 0.65
query7 0.03 0.02 0.02
query8 0.05 0.04 0.04
query9 0.58 0.51 0.50
query10 0.56 0.57 0.56
query11 0.16 0.11 0.12
query12 0.14 0.11 0.12
query13 0.63 0.60 0.60
query14 0.99 0.98 0.98
query15 0.82 0.82 0.81
query16 0.40 0.41 0.39
query17 1.04 0.99 1.02
query18 0.23 0.22 0.21
query19 1.88 1.86 1.83
query20 0.02 0.01 0.01
query21 15.45 0.29 0.14
query22 4.92 0.05 0.04
query23 16.04 0.30 0.11
query24 1.27 0.60 0.22
query25 0.09 0.06 0.07
query26 0.14 0.14 0.13
query27 0.08 0.05 0.05
query28 4.92 1.21 1.02
query29 12.57 3.98 3.20
query30 0.29 0.14 0.11
query31 2.83 0.61 0.40
query32 3.23 0.56 0.47
query33 3.05 3.06 3.05
query34 16.91 5.19 4.50
query35 4.49 4.56 4.50
query36 0.69 0.51 0.47
query37 0.11 0.07 0.07
query38 0.08 0.04 0.04
query39 0.05 0.03 0.03
query40 0.17 0.14 0.14
query41 0.09 0.03 0.03
query42 0.04 0.03 0.03
query43 0.04 0.03 0.04
Total cold run time: 98.52 s
Total hot run time: 27.21 s
```
</details>
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3602968523
<details>
<summary>TPC-H: <b>Total hot run time: 34030 ms</b></summary>
```
machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
scripts: https://github.com/apache/doris/tree/master/tools/tpch-tools
Tpch sf100 test result on commit 1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113, data reload: false
------ Round 1 ----------------------------------
q1 17620 5059 4947 4947
q2 2029 308 196 196
q3 10265 1287 726 726
q4 10209 818 315 315
q5 7536 2398 2130 2130
q6 183 169 139 139
q7 939 802 648 648
q8 9342 1417 1057 1057
q9 6880 5255 5274 5255
q10 6817 2197 1783 1783
q11 509 309 291 291
q12 336 365 226 226
q13 17775 3702 3003 3003
q14 226 228 216 216
q15 583 504 507 504
q16 895 871 825 825
q17 675 793 514 514
q18 7812 7160 6963 6963
q19 1094 939 581 581
q20 363 357 228 228
q21 4008 3095 2550 2550
q22 1020 969 933 933
Total cold run time: 107116 ms
Total hot run time: 34030 ms
----- Round 2, with runtime_filter_mode=off -----
q1 5009 4978 5026 4978
q2 325 400 311 311
q3 2138 2664 2279 2279
q4 1323 1718 1260 1260
q5 4222 4380 4436 4380
q6 229 179 138 138
q7 2098 2030 1832 1832
q8 2737 2611 2518 2518
q9 7529 7545 7398 7398
q10 3102 3294 2817 2817
q11 599 544 495 495
q12 684 741 604 604
q13 3625 4086 3267 3267
q14 278 289 272 272
q15 548 510 494 494
q16 912 932 870 870
q17 1210 1431 1428 1428
q18 7868 7732 7521 7521
q19 852 845 911 845
q20 2145 2094 1872 1872
q21 4798 4230 4224 4224
q22 1112 1026 990 990
Total cold run time: 53343 ms
Total hot run time: 50793 ms
```
</details>
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "kevinw66 (via GitHub)" <gi...@apache.org>.
kevinw66 commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3604802818
run buildall
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3604952337
# BE UT Coverage Report
Increment line coverage `100.00% (15/15)` :tada:
[Increment coverage report](http://coverage.selectdb-in.cc/coverage/1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113/increment_report/index.html)
[Complete coverage report](http://coverage.selectdb-in.cc/coverage/1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113/report/index.html)
| Category | Coverage |
|-------------------|--------------------|
| Function Coverage | 53.51% (18734/35010) |
| Line Coverage | 39.10% (172708/441708) |
| Region Coverage | 33.69% (133647/396689) |
| Branch Coverage | 34.65% (57553/166082) |
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
Re: [PR] [opt](arm) Improve count_zero_num performance with NEON intrinsics [doris]
Posted by "hello-stephen (via GitHub)" <gi...@apache.org>.
hello-stephen commented on PR #58615:
URL: https://github.com/apache/doris/pull/58615#issuecomment-3606084567
# BE Regression && UT Coverage Report
Increment line coverage `100.00% (15/15)` :tada:
[Increment coverage report](http://coverage.selectdb-in.cc/coverage/58615_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_merge/increment_report/index.html)
[Complete coverage report](http://coverage.selectdb-in.cc/coverage/58615_1233a0f33fb3ef95f13ecad392ea1e6f5f6d3113_merge/report/index.html)
| Category | Coverage |
|-------------------|--------------------|
| Function Coverage | 72.32% (24823/34326) |
| Line Coverage | 58.95% (260156/441313) |
| Region Coverage | 53.78% (215910/401499) |
| Branch Coverage | 55.38% (92467/166979) |
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org