ODIn/spark-instrumented-optimizer
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity
spark-instrumented-optimizer/sql/hive/benchmarks/OrcReadBenchmark-results.txt

157 lines
13 KiB
Plaintext
Raw Normal View History

[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
================================================================================================
SQL Single Numeric Column Scan
================================================================================================
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
SQL Single TINYINT Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 1843 1958 162 8.5 117.2 1.0X
Native ORC Vectorized 321 355 31 48.9 20.4 5.7X
Hive built-in ORC 2143 2175 44 7.3 136.3 0.9X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
SQL Single SMALLINT Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 1987 2020 47 7.9 126.3 1.0X
Native ORC Vectorized 276 299 25 57.0 17.6 7.2X
Hive built-in ORC 2350 2357 10 6.7 149.4 0.8X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
SQL Single INT Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 2092 2115 32 7.5 133.0 1.0X
Native ORC Vectorized 360 373 18 43.6 22.9 5.8X
Hive built-in ORC 2550 2557 9 6.2 162.2 0.8X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
SQL Single BIGINT Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 2173 2188 21 7.2 138.2 1.0X
Native ORC Vectorized 435 448 14 36.2 27.7 5.0X
Hive built-in ORC 2683 2690 10 5.9 170.6 0.8X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
SQL Single FLOAT Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 2233 2323 127 7.0 142.0 1.0X
Native ORC Vectorized 475 483 13 33.1 30.2 4.7X
Hive built-in ORC 2605 2610 6 6.0 165.7 0.9X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
SQL Single DOUBLE Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 2367 2384 24 6.6 150.5 1.0X
Native ORC Vectorized 600 641 69 26.2 38.1 3.9X
Hive built-in ORC 2860 2877 24 5.5 181.9 0.8X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
================================================================================================
Int and String Scan
================================================================================================
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
Int and String Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 4253 4330 108 2.5 405.6 1.0X
Native ORC Vectorized 2295 2301 8 4.6 218.9 1.9X
Hive built-in ORC 5364 5465 144 2.0 511.5 0.8X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
================================================================================================
Partitioned Table Scan
================================================================================================
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
Partitioned Table: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Data column - Native ORC MR 2443 2448 6 6.4 155.3 1.0X
Data column - Native ORC Vectorized 446 473 44 35.3 28.3 5.5X
Data column - Hive built-in ORC 2868 2877 12 5.5 182.4 0.9X
Partition column - Native ORC MR 1623 1656 47 9.7 103.2 1.5X
Partition column - Native ORC Vectorized 112 121 14 140.8 7.1 21.9X
Partition column - Hive built-in ORC 1846 1850 5 8.5 117.4 1.3X
Both columns - Native ORC MR 2610 2635 36 6.0 165.9 0.9X
Both columns - Native ORC Vectorized 492 508 19 32.0 31.3 5.0X
Both columns - Hive built-in ORC 2969 2973 4 5.3 188.8 0.8X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
================================================================================================
Repeated String Scan
================================================================================================
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
Repeated String: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 2056 2064 11 5.1 196.1 1.0X
Native ORC Vectorized 415 421 7 25.3 39.6 5.0X
Hive built-in ORC 2710 2722 17 3.9 258.4 0.8X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
================================================================================================
String with Nulls Scan
================================================================================================
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
String with Nulls Scan (0.0%): Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 3655 3674 27 2.9 348.6 1.0X
Native ORC Vectorized 1166 1167 1 9.0 111.2 3.1X
Hive built-in ORC 5268 5305 52 2.0 502.4 0.7X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
String with Nulls Scan (50.0%): Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 3447 3467 27 3.0 328.8 1.0X
Native ORC Vectorized 1222 1223 1 8.6 116.6 2.8X
Hive built-in ORC 3947 3959 18 2.7 376.4 0.9X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
String with Nulls Scan (95.0%): Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 1912 1917 6 5.5 182.4 1.0X
Native ORC Vectorized 477 484 5 22.0 45.5 4.0X
Hive built-in ORC 2374 2386 17 4.4 226.4 0.8X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
================================================================================================
Single Column Scan From Wide Columns
================================================================================================
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
Single Column Scan from 100 columns: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 290 350 102 3.6 276.1 1.0X
Native ORC Vectorized 155 166 15 6.7 148.2 1.9X
Hive built-in ORC 520 531 8 2.0 495.8 0.6X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
Single Column Scan from 200 columns: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 365 406 73 2.9 347.9 1.0X
Native ORC Vectorized 232 246 20 4.5 221.6 1.6X
Hive built-in ORC 794 864 62 1.3 757.6 0.5X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
OpenJDK 64-Bit Server VM 1.8.0_222-b10 on Linux 3.10.0-862.3.2.el7.x86_64
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29141][SQL][TEST] Use SqlBasedBenchmark in SQL benchmarks ### What changes were proposed in this pull request? Refactored SQL-related benchmark and made them depend on `SqlBasedBenchmark`. In particular, creation of Spark session are moved into `override def getSparkSession: SparkSession`. ### Why are the changes needed? This should simplify maintenance of SQL-based benchmarks by reducing the number of dependencies. In the future, it should be easier to refactor & extend all SQL benchmarks by changing only one trait. Finally, all SQL-based benchmarks will look uniformly. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the modified benchmarks. Closes #25828 from MaxGekk/sql-benchmarks-refactoring. Lead-authored-by: Maxim Gekk <max.gekk@gmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
2019-09-18 20:52:23 -04:00
Single Column Scan from 300 columns: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Native ORC MR 501 544 40 2.1 477.6 1.0X
Native ORC Vectorized 365 386 33 2.9 348.0 1.4X
Hive built-in ORC 1153 1153 0 0.9 1099.8 0.4X
[SPARK-25508][SQL][TEST] Refactor OrcReadBenchmark to use main method ## What changes were proposed in this pull request? Refactor OrcReadBenchmark to use main method. Generate benchmark result: ``` SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "hive/test:runMain org.apache.spark.sql.hive.orc.OrcReadBenchmark" ``` ## How was this patch tested? manual tests Closes #22580 from yucai/SPARK-25508. Lead-authored-by: yucai <yyu1@ebay.com> Co-authored-by: Yucai Yu <yucai.yu@foxmail.com> Co-authored-by: Dongjoon Hyun <dongjoon@apache.org> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2018-09-29 12:48:03 -04:00
Reference in a new issue Copy permalink