huaxingao commented on code in PR #36218: URL: https://github.com/apache/spark/pull/36218#discussion_r851671963
########## sql/core/benchmarks/BloomFilterBenchmark-jdk17-results.txt: ########## @@ -3,118 +3,190 @@ ORC Write ================================================================================================ OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure -Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz Write 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative ------------------------------------------------------------------------------------------------------------------------ -Without bloom filter 19097 19523 602 5.2 191.0 1.0X -With bloom filter 22213 22402 267 4.5 222.1 0.9X +Without bloom filter 15254 15831 816 6.6 152.5 1.0X +With bloom filter 18586 18608 32 5.4 185.9 0.8X ================================================================================================ ORC Read ================================================================================================ OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure -Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative ------------------------------------------------------------------------------------------------------------------------ -Without bloom filter 1528 1628 142 65.4 15.3 1.0X -With bloom filter 1370 1391 29 73.0 13.7 1.1X +Without bloom filter, blocksize: 2097152 1199 1225 37 83.4 12.0 1.0X +With bloom filter, blocksize: 2097152 886 894 11 112.8 8.9 1.4X + + +================================================================================================ +ORC Read +================================================================================================ + +OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz +Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------------------------------ +Without bloom filter, blocksize: 4194304 1202 1208 8 83.2 12.0 1.0X +With bloom filter, blocksize: 4194304 874 893 30 114.4 8.7 1.4X + + +================================================================================================ +ORC Read +================================================================================================ + +OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz +Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------------------------------ +Without bloom filter, blocksize: 6291456 1241 1251 15 80.6 12.4 1.0X +With bloom filter, blocksize: 6291456 852 857 6 117.4 8.5 1.5X + + +================================================================================================ +ORC Read +================================================================================================ + +OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz +Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------------------------------ +Without bloom filter, blocksize: 8388608 1182 1183 2 84.6 11.8 1.0X +With bloom filter, blocksize: 8388608 852 859 7 117.3 8.5 1.4X + + +================================================================================================ +ORC Read +================================================================================================ + +OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz +Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------------------------------- +Without bloom filter, blocksize: 12582912 1185 1187 2 84.4 11.8 1.0X +With bloom filter, blocksize: 12582912 857 883 23 116.6 8.6 1.4X + + +================================================================================================ +ORC Read +================================================================================================ + +OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz +Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------------------------------- +Without bloom filter, blocksize: 16777216 1194 1197 5 83.8 11.9 1.0X +With bloom filter, blocksize: 16777216 837 848 10 119.4 8.4 1.4X + + +================================================================================================ +ORC Read +================================================================================================ + +OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz +Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------------------------------- +Without bloom filter, blocksize: 33554432 1149 1151 4 87.0 11.5 1.0X +With bloom filter, blocksize: 33554432 841 850 8 119.0 8.4 1.4X ================================================================================================ Parquet Write ================================================================================================ OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure -Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz Write 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative ------------------------------------------------------------------------------------------------------------------------ -Without bloom filter 17846 17981 191 5.6 178.5 1.0X -With bloom filter 22883 22982 140 4.4 228.8 0.8X +Without bloom filter 16009 16144 191 6.2 160.1 1.0X +With bloom filter 19576 19850 387 5.1 195.8 0.8X ================================================================================================ Parquet Read ================================================================================================ OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure -Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative ------------------------------------------------------------------------------------------------------------------------ -Without bloom filter, blocksize: 2097152 947 1012 92 105.6 9.5 1.0X -With bloom filter, blocksize: 2097152 311 325 12 322.0 3.1 3.0X +Without bloom filter, blocksize: 2097152 712 754 43 140.5 7.1 1.0X +With bloom filter, blocksize: 2097152 227 236 9 440.9 2.3 3.1X ================================================================================================ Parquet Read ================================================================================================ OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure -Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative ------------------------------------------------------------------------------------------------------------------------ -Without bloom filter, blocksize: 4194304 824 837 15 121.3 8.2 1.0X -With bloom filter, blocksize: 4194304 210 222 10 475.8 2.1 3.9X +Without bloom filter, blocksize: 4194304 674 677 3 148.3 6.7 1.0X +With bloom filter, blocksize: 4194304 160 167 5 625.3 1.6 4.2X ================================================================================================ Parquet Read ================================================================================================ OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure -Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative ------------------------------------------------------------------------------------------------------------------------ -Without bloom filter, blocksize: 6291456 1210 1327 166 82.6 12.1 1.0X -With bloom filter, blocksize: 6291456 302 315 9 331.1 3.0 4.0X +Without bloom filter, blocksize: 6291456 675 680 7 148.2 6.7 1.0X +With bloom filter, blocksize: 6291456 185 190 5 540.9 1.8 3.6X ================================================================================================ Parquet Read ================================================================================================ OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure -Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative ------------------------------------------------------------------------------------------------------------------------ -Without bloom filter, blocksize: 8388608 1244 1251 10 80.4 12.4 1.0X -With bloom filter, blocksize: 8388608 504 513 8 198.4 5.0 2.5X +Without bloom filter, blocksize: 8388608 673 695 28 148.5 6.7 1.0X +With bloom filter, blocksize: 8388608 306 310 3 327.2 3.1 2.2X ================================================================================================ Parquet Read ================================================================================================ OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure -Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative ------------------------------------------------------------------------------------------------------------------------- -Without bloom filter, blocksize: 12582912 1384 1408 34 72.3 13.8 1.0X -With bloom filter, blocksize: 12582912 900 955 49 111.1 9.0 1.5X +Without bloom filter, blocksize: 12582912 678 681 4 147.4 6.8 1.0X +With bloom filter, blocksize: 12582912 497 501 5 201.0 5.0 1.4X ================================================================================================ Parquet Read ================================================================================================ OpenJDK 64-Bit Server VM 17.0.2+8-LTS on Linux 5.13.0-1021-azure -Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz +Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz Read a row from 100M rows: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative ------------------------------------------------------------------------------------------------------------------------- -Without bloom filter, blocksize: 16777216 1468 1482 19 68.1 14.7 1.0X -With bloom filter, blocksize: 16777216 1302 1313 15 76.8 13.0 1.1X +Without bloom filter, blocksize: 16777216 675 738 102 148.2 6.7 1.0X +With bloom filter, blocksize: 16777216 1071 1075 5 93.4 10.7 0.6X Review Comment: Shall we rerun this to see if the anomaly will go away? -- 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]
