You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by GitBox <gi...@apache.org> on 2022/01/14 18:54:08 UTC
[GitHub] [spark] dongjoon-hyun commented on a change in pull request #35212: [SPARK-36879][SQL][TESTS] Add benchmark for vectorized delta binary packed encoding
dongjoon-hyun commented on a change in pull request #35212:
URL: https://github.com/apache/spark/pull/35212#discussion_r785079082
##########
File path: sql/core/benchmarks/DataSourceReadBenchmark-jdk11-results.txt
##########
@@ -2,269 +2,283 @@
SQL Single Numeric Column Scan
================================================================================================
-OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1022-azure
-Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz
+OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1025-azure
+Intel(R) Xeon(R) CPU E5-2673 v4 @ 2.30GHz
SQL Single BOOLEAN Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
-SQL CSV 11834 11929 134 1.3 752.4 1.0X
-SQL Json 8574 8597 32 1.8 545.1 1.4X
-SQL Parquet Vectorized 116 136 17 135.5 7.4 102.0X
-SQL Parquet MR 1703 1715 17 9.2 108.2 7.0X
-SQL ORC Vectorized 172 215 48 91.2 11.0 68.6X
-SQL ORC MR 1819 1825 8 8.6 115.7 6.5X
-
-OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1022-azure
-Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz
+SQL CSV 12783 12908 176 1.2 812.8 1.0X
+SQL Json 10235 10252 24 1.5 650.8 1.2X
+SQL Parquet Vectorized 129 162 21 122.1 8.2 99.2X
+SQL Parquet MR 2170 2185 20 7.2 138.0 5.9X
+SQL Parquet Vectorized (Delta Binary) 94 129 34 166.8 6.0 135.6X
+SQL Parquet MR (Delta Binary) 2051 2056 8 7.7 130.4 6.2X
+SQL ORC Vectorized 177 234 47 88.8 11.3 72.2X
+SQL ORC MR 1589 1606 24 9.9 101.0 8.0X
+
+OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1025-azure
+Intel(R) Xeon(R) CPU E5-2673 v4 @ 2.30GHz
Parquet Reader Single BOOLEAN Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
-------------------------------------------------------------------------------------------------------------------------
-ParquetReader Vectorized 117 126 17 134.9 7.4 1.0X
-ParquetReader Vectorized -> Row 47 49 3 336.5 3.0 2.5X
+ParquetReader Vectorized 99 115 15 158.8 6.3 1.0X
+ParquetReader Vectorized -> Row 46 51 4 345.0 2.9 2.2X
-OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1022-azure
-Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz
+OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1025-azure
+Intel(R) Xeon(R) CPU E5-2673 v4 @ 2.30GHz
SQL Single TINYINT Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
-SQL CSV 13434 13590 220 1.2 854.1 1.0X
-SQL Json 10056 10073 24 1.6 639.3 1.3X
-SQL Parquet Vectorized 212 229 12 74.3 13.5 63.4X
-SQL Parquet MR 1883 1916 47 8.4 119.7 7.1X
-SQL ORC Vectorized 200 241 30 78.8 12.7 67.3X
-SQL ORC MR 1529 1549 28 10.3 97.2 8.8X
-
-OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1022-azure
-Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz
+SQL CSV 15867 15952 120 1.0 1008.8 1.0X
+SQL Json 11170 11203 47 1.4 710.2 1.4X
+SQL Parquet Vectorized 147 174 39 106.7 9.4 107.6X
+SQL Parquet MR 2352 2361 12 6.7 149.6 6.7X
+SQL Parquet Vectorized (Delta Binary) 146 172 27 108.0 9.3 109.0X
+SQL Parquet MR (Delta Binary) 2139 2169 43 7.4 136.0 7.4X
+SQL ORC Vectorized 175 225 51 90.1 11.1 90.9X
+SQL ORC MR 1566 1588 31 10.0 99.6 10.1X
+
+OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1025-azure
+Intel(R) Xeon(R) CPU E5-2673 v4 @ 2.30GHz
Parquet Reader Single TINYINT Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
-------------------------------------------------------------------------------------------------------------------------
-ParquetReader Vectorized 229 254 13 68.6 14.6 1.0X
-ParquetReader Vectorized -> Row 162 171 14 96.9 10.3 1.4X
+ParquetReader Vectorized 175 190 13 89.8 11.1 1.0X
+ParquetReader Vectorized -> Row 96 108 17 163.6 6.1 1.8X
-OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1022-azure
-Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz
+OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1025-azure
+Intel(R) Xeon(R) CPU E5-2673 v4 @ 2.30GHz
SQL Single SMALLINT Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
-SQL CSV 14320 14476 221 1.1 910.4 1.0X
-SQL Json 9769 10067 423 1.6 621.1 1.5X
-SQL Parquet Vectorized 187 228 28 84.3 11.9 76.8X
-SQL Parquet MR 2230 2240 14 7.1 141.8 6.4X
-SQL ORC Vectorized 221 265 36 71.1 14.1 64.8X
-SQL ORC MR 1763 1779 23 8.9 112.1 8.1X
-
-OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1022-azure
-Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz
+SQL CSV 16787 17659 1234 0.9 1067.3 1.0X
+SQL Json 11373 11520 209 1.4 723.1 1.5X
+SQL Parquet Vectorized 205 259 26 76.6 13.0 81.8X
+SQL Parquet MR 2673 2698 35 5.9 169.9 6.3X
+SQL Parquet Vectorized (Delta Binary) 252 291 38 62.5 16.0 66.7X
+SQL Parquet MR (Delta Binary) 2220 2280 85 7.1 141.1 7.6X
+SQL ORC Vectorized 276 295 30 57.1 17.5 60.9X
+SQL ORC MR 1835 1852 25 8.6 116.6 9.2X
+
+OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1025-azure
+Intel(R) Xeon(R) CPU E5-2673 v4 @ 2.30GHz
Parquet Reader Single SMALLINT Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
--------------------------------------------------------------------------------------------------------------------------
-ParquetReader Vectorized 246 255 12 64.1 15.6 1.0X
-ParquetReader Vectorized -> Row 249 294 21 63.1 15.8 1.0X
+ParquetReader Vectorized 255 276 19 61.7 16.2 1.0X
+ParquetReader Vectorized -> Row 285 307 15 55.1 18.1 0.9X
-OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1022-azure
-Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz
+OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1025-azure
+Intel(R) Xeon(R) CPU E5-2673 v4 @ 2.30GHz
SQL Single INT Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
-SQL CSV 15460 15543 116 1.0 982.9 1.0X
-SQL Json 10199 10393 274 1.5 648.4 1.5X
-SQL Parquet Vectorized 163 203 30 96.5 10.4 94.8X
-SQL Parquet MR 1914 2025 157 8.2 121.7 8.1X
-SQL ORC Vectorized 324 355 23 48.5 20.6 47.7X
-SQL ORC MR 1673 1701 39 9.4 106.4 9.2X
-
-OpenJDK 64-Bit Server VM 11.0.13+8-LTS on Linux 5.11.0-1022-azure
-Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz
+SQL CSV 18402 19372 1373 0.9 1169.9 1.0X
+SQL Json 11883 12025 202 1.3 755.5 1.5X
+SQL Parquet Vectorized 228 244 9 69.0 14.5 80.8X
+SQL Parquet MR 2616 2630 19 6.0 166.3 7.0X
+SQL Parquet Vectorized (Delta Binary) 463 510 53 34.0 29.4 39.8X
Review comment:
Ditto. Does this mean a regression from `80x` to `40x`? Two times slower than before?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org