msfroh commented on PR #13054: URL: https://github.com/apache/lucene/pull/13054#issuecomment-1935491971
I decided to try experimenting with moving the output words back onto the heap, since I didn't the fact that every word lookup was triggering a seek. Running now, I got way less variance on the on-heap runs. I also added some GCs between iterations, since I wanted to measure the heap usage of each. That likely removed some GC pauses from the on-heap version. I then switched back to the off-heap words to confirm the results that I saw last time (and compare against the implementation with on-heap words). The conclusion seems to be roughly: * Existing on-heap FST averages about 444ms to process a lot of synonyms. * Off-heap FST with on-heap words averages 515 or 516ms. (About 16% slower than existing on-heap.) * Off-heap FST with off-heap words averages 620ms. (About 40% slower than existing on-heap.) The on-heap FST seems to occupy about 36MB of heap. The off-heap FST with on-heap words occupies about 560kB. The off-heap FST with off-heap words occupies about 150kB. With these trade-offs, I think off-heap FST with on-heap words may be a good choice for folks with large sets of synonyms. I don't think I would recommend off-heap FST with off-heap words. | Attempt OnHeap FST load time | OffHeap FST (OnHeap words) load time | OffHeap FST (OnHeap words) reload time | OnHeap FST processing time | OffHeap FST (OnHeap words) processing time | OffHeap FST (OnHeap words) reloaded processing time | OffHeap FST (OffHeap words) processing time | OffHeap FST (OffHeap words) reloaded processing time |------------------------------|--------------------------------------|----------------------------------------|----------------------------|--------------------------------------------|-----------------------------------------------------|---------------------------------------------|----------------------------------------------------- | 1 | 1191.339685 | 1072.285824 | 9.669646 | 436.391631 | 520.550704 | 516.11297 | 623.451546 | 620.531215 | 2 | 1030.432454 | 1033.619768 | 8.874105 | 448.848403 | 516.784387 | 517.230739 | 621.522464 | 622.793343 | 3 | 984.83645 | 1037.807342 | 8.912252 | 443.789813 | 512.066535 | 517.716981 | 622.455444 | 620.468985 | 4 | 1049.63589 | 1048.60113 | 8.894401 | 449.237547 | 518.946226 | 516.868933 | 617.837364 | 616.810236 | 5 | 990.22176 | 1049.618665 | 8.861166 | 448.923912 | 512.559801 | 511.114898 | 616.555422 | 617.122551 | 6 | 978.41877 | 1063.824595 | 8.930418 | 440.251675 | 517.632376 | 518.175232 | 621.969759 | 622.828416 | 7 | 985.434177 | 1049.113913 | 8.872906 | 443.209607 | 511.210536 | 518.802292 | 624.151468 | 622.097039 | 8 | 985.376238 | 1046.102696 | 8.823786 | 440.815454 | 517.491411 | 517.905752 | 623.390319 | 625.387487 | 9 | 983.341325 | 1065.892279 | 8.871586 | 449.145252 | 516.029267 | 516.916524 | 622.811992 | 622.798858 | 10 | 985.438642 | 1046.71167 | 8.8518 | 445.970679 | 512.045037 | 518.934149 | 622.592098 | 614.661805 | 11 | 990.592624 | 1050.377106 | 8.832753 | 443.844237 | 515.758106 | 510.808005 | 611.62254 | 622.956946 | 12 | 986.747374 | 1066.052969 | 8.884928 | 444.398327 | 517.259451 | 524.770132 | 622.085785 | 619.311172 | 13 | 984.328191 | 1052.189621 | 8.88281 | 439.612497 | 517.861131 | 515.796013 | 617.862222 | 615.101452 | 14 | 984.405339 | 1049.06783 | 8.835775 | 438.871305 | 517.885493 | 515.853446 | 615.254987 | 623.464483 | 15 | 997.323593 | 1064.473985 | 8.90682 | 443.640208 | 515.329143 | 518.807239 | 623.020916 | 623.013801 | 16 | 997.253932 | 1066.558928 | 8.900308 | 442.534843 | 511.930766 | 516.365803 | 624.316916 | 615.037306 | 17 | 999.464751 | 1046.464149 | 8.895899 | 443.48306 | 514.841946 | 517.082166 | 617.615908 | 618.661376 | 18 | 1001.896073 | 1045.304622 | 8.877555 | 444.875225 | 515.029862 | 510.365428 | 618.540866 | 624.355309 | 19 | 986.055833 | 1045.208347 | 8.863339 | 441.647553 | 511.489699 | 517.213428 | 623.61503 | 621.198543 | 20 | 984.112667 | 1047.317164 | 8.940865 | 451.304206 | 514.762544 | 510.45981 | 621.057397 | 621.483146 | 21 | 988.310511 | 1046.154648 | 8.865301 | 447.25874 | 514.859414 | 517.24163 | 623.916511 | 614.185296 | 22 | 982.874582 | 1062.113889 | 8.867098 | 439.785463 | 510.387721 | 516.885653 | 623.494968 | 622.527091 | 23 | 980.96967 | 1048.050631 | 8.867966 | 439.05464 | 511.423329 | 516.984465 | 621.567988 | 621.204435 | 24 | 983.189843 | 1046.083632 | 8.81578 | 440.574651 | 518.390122 | 520.392926 | 622.34785 | 614.923018 | 25 | 987.033178 | 1074.553767 | 8.812579 | 446.687106 | 513.914686 | 521.952744 | 615.870183 | 621.089011 | 26 | 985.771758 | 1076.245942 | 8.845264 | 444.718264 | 516.274395 | 513.5547 | 615.927497 | 615.53522 | 27 | 981.748774 | 1046.85677 | 8.818164 | 443.252924 | 513.632714 | 515.919924 | 626.659516 | 622.307368 | 28 | 983.979894 | 1062.317764 | 8.869256 | 443.267803 | 513.965345 | 509.688356 | 615.790469 | 615.712761 | 29 | 980.908776 | 1045.006602 | 8.855109 | 444.452376 | 517.488159 | 509.770143 | 621.96871 | 621.582871 | 30 | 981.508232 | 1046.722776 | 8.790313 | 443.952753 | 513.840793 | 512.847346 | 621.747601 | 621.901271 | 31 | 999.165558 | 1063.517734 | 8.792905 | 440.356205 | 517.677777 | 517.920992 | 620.90204 | 613.422668 | 32 | 1000.854281 | 1060.766663 | 9.027399 | 444.385706 | 510.006231 | 514.006688 | 623.684492 | 620.742008 | 33 | 1001.620724 | 1046.329083 | 8.72687 | 443.912072 | 509.793229 | 513.313214 | 620.695915 | 621.266234 | 34 | 1008.677463 | 1044.437966 | 8.799494 | 447.077333 | 516.263674 | 514.751767 | 622.775084 | 620.885167 | 35 | 987.309353 | 1048.062722 | 8.763122 | 440.748052 | 518.972785 | 518.608101 | 621.032898 | 620.85482 | 36 | 980.960836 | 1052.037316 | 8.834358 | 445.210623 | 518.850346 | 511.742763 | 620.719135 | 621.679027 | 37 | 983.807955 | 1049.894433 | 8.798039 | 440.302584 | 511.351473 | 510.557417 | 615.059624 | 619.802549 | 38 | 982.144377 | 1071.744423 | 8.81421 | 444.036711 | 518.551589 | 515.779265 | 614.579103 | 615.092139 | 39 | 984.460101 | 1051.399337 | 8.781058 | 439.998112 | 518.709639 | 511.192122 | 614.797646 | 621.154429 | 40 | 981.16924 | 1047.739329 | 8.827411 | 446.515425 | 512.815557 | 519.138446 | 621.769829 | 613.83963 | **Mean** | **995.5780219** | **1053.415701** | **8.87387035** | **443.6585744** | **515.115835** | **515.7387151** | **620.4259376** | **619.7447621** | StdDev | 34.59108879 | 10.41692769 | 0.139732265 | 3.344053675 | 2.951489941 | 3.528886642 | 3.48797555 | 3.346472427 -- 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: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
