msokolov commented on PR #16805: URL: https://github.com/apache/lucene/pull/16805#issuecomment-6619063776
with this change I got these benchmark results: ## BEFORE main (16 searcher threads) ``` ------------> Report Sum By (any) Name (5 about 13 out of 13) Operation round runCnt recsPerRun rec/s elapsedSec avgUsedMem avgTotalMem Rounds_3 0 1 6000 3.82 1,571.62 250,299,104 534,773,760 term_500 - - - - - - 3 - - - 500 - - - 1.74 - - 864.38 - 176,789,840 - 533,725,184 numeric_500 - 3 500 2.34 641.23 207,488,576 533,725,184 globalOrd_500 - - - - 3 - - - 500 - - 25.72 - - 58.33 - 273,694,240 - 534,074,720 block_500 - 3 500 195.26 7.68 219,027,584 534,074,720 ``` ## AFTER collector-manager-join 16 threads ``` ------------> Report Sum By (any) Name (5 about 13 out of 13) Operation round runCnt recsPerRun rec/s elapsedSec avgUsedMem avgTotalMem Rounds_3 0 1 6000 4.08 1,468.85 267,262,528 533,725,184 term_500 - - - - - - 3 - - - 500 - - - 1.94 - - 774.51 - 191,378,288 - 533,725,184 numeric_500 - 3 500 2.39 628.84 167,292,352 533,725,184 globalOrd_500 - - - - 3 - - - 500 - - 25.95 - - 57.80 - 411,140,448 - 533,725,184 block_500 - 3 500 194.73 7.70 411,140,448 533,725,184 ``` seems maybe slightly faster. Anyway no regression -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
