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

41 lines
4.3 KiB
Plaintext
Raw Normal View History

[SPARK-32071][SQL][TESTS] Add `make_interval` benchmark ### What changes were proposed in this pull request? Add benchmarks for interval constructor `make_interval` and measure perf of 4 cases: 1. Constant (year, month) 2. Constant (week, day) 3. Constant (hour, minute, second, second fraction) 4. All fields are NOT constant. The benchmark results are generated in the environment: | Item | Description | | ---- | ----| | Region | us-west-2 (Oregon) | | Instance | r3.xlarge | | AMI | ubuntu/images/hvm-ssd/ubuntu-bionic-18.04-amd64-server-20190722.1 (ami-06f2f779464715dc5) | | Java | OpenJDK 64-Bit Server VM 1.8.0_252 and OpenJDK 64-Bit Server VM 11.0.7+10 | ### Why are the changes needed? To have a base line for future perf improvements of `make_interval`, and to prevent perf regressions in the future. ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? By running `IntervalBenchmark` via: ``` $ SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "sql/test:runMain org.apache.spark.sql.execution.benchmark.IntervalBenchmark" ``` Closes #28905 from MaxGekk/benchmark-make_interval. Authored-by: Max Gekk <max.gekk@gmail.com> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2020-06-27 20:54:06 -04:00
OpenJDK 64-Bit Server VM 1.8.0_252-8u252-b09-1~18.04-b09 on Linux 4.15.0-1063-aws
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
[SPARK-29533][SQL][TEST] Benchmark casting strings to intervals ### What changes were proposed in this pull request? Added new benchmark `IntervalBenchmark` to measure performance of interval related functions. In the PR, I added benchmarks for casting strings to interval. In particular, interval strings with `interval` prefix and without it because there is special code for this https://github.com/apache/spark/blob/da576a737c2db01e5ba5ce19ed0e8f900cb5efaf/common/unsafe/src/main/java/org/apache/spark/unsafe/types/CalendarInterval.java#L100-L103 . And also I added benchmarks for different number of units in interval strings, for example 1 unit is `interval 10 years`, 2 units w/o interval is `10 years 5 months`, and etc. ### Why are the changes needed? - To find out current performance issues in casting to intervals - The benchmark can be used while refactoring/re-implementing `CalendarInterval.fromString()` or `CalendarInterval.fromCaseInsensitiveString()`. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the benchmark via the command: ```shell SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "sql/test:runMain org.apache.spark.sql.execution.benchmark.IntervalBenchmark" ``` Closes #26189 from MaxGekk/interval-from-string-benchmark. Authored-by: Maxim Gekk <max.gekk@gmail.com> Signed-off-by: HyukjinKwon <gurwls223@apache.org>
2019-10-21 21:47:04 -04:00
cast strings to intervals: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
[SPARK-32071][SQL][TESTS] Add `make_interval` benchmark ### What changes were proposed in this pull request? Add benchmarks for interval constructor `make_interval` and measure perf of 4 cases: 1. Constant (year, month) 2. Constant (week, day) 3. Constant (hour, minute, second, second fraction) 4. All fields are NOT constant. The benchmark results are generated in the environment: | Item | Description | | ---- | ----| | Region | us-west-2 (Oregon) | | Instance | r3.xlarge | | AMI | ubuntu/images/hvm-ssd/ubuntu-bionic-18.04-amd64-server-20190722.1 (ami-06f2f779464715dc5) | | Java | OpenJDK 64-Bit Server VM 1.8.0_252 and OpenJDK 64-Bit Server VM 11.0.7+10 | ### Why are the changes needed? To have a base line for future perf improvements of `make_interval`, and to prevent perf regressions in the future. ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? By running `IntervalBenchmark` via: ``` $ SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "sql/test:runMain org.apache.spark.sql.execution.benchmark.IntervalBenchmark" ``` Closes #28905 from MaxGekk/benchmark-make_interval. Authored-by: Max Gekk <max.gekk@gmail.com> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
2020-06-27 20:54:06 -04:00
prepare string w/ interval 677 718 40 1.5 677.2 1.0X
prepare string w/o interval 602 624 19 1.7 602.2 1.1X
1 units w/ interval 582 598 20 1.7 581.8 1.2X
1 units w/o interval 549 591 64 1.8 549.1 1.2X
2 units w/ interval 758 773 14 1.3 758.2 0.9X
2 units w/o interval 723 738 14 1.4 722.6 0.9X
3 units w/ interval 1442 1450 11 0.7 1441.8 0.5X
3 units w/o interval 1426 1429 3 0.7 1426.4 0.5X
4 units w/ interval 1645 1652 11 0.6 1645.1 0.4X
4 units w/o interval 1618 1626 10 0.6 1617.6 0.4X
5 units w/ interval 1794 1803 13 0.6 1794.4 0.4X
5 units w/o interval 1783 1793 9 0.6 1783.2 0.4X
6 units w/ interval 1976 1984 11 0.5 1976.2 0.3X
6 units w/o interval 1948 1959 10 0.5 1947.9 0.3X
7 units w/ interval 2394 2408 18 0.4 2393.7 0.3X
7 units w/o interval 2387 2392 8 0.4 2386.8 0.3X
8 units w/ interval 2578 2588 15 0.4 2577.5 0.3X
8 units w/o interval 2572 2578 5 0.4 2571.8 0.3X
9 units w/ interval 2812 2829 19 0.4 2811.7 0.2X
9 units w/o interval 2811 2816 4 0.4 2810.7 0.2X
10 units w/ interval 3108 3116 10 0.3 3107.8 0.2X
10 units w/o interval 3107 3109 3 0.3 3106.8 0.2X
11 units w/ interval 3386 3392 8 0.3 3386.3 0.2X
11 units w/o interval 3374 3377 4 0.3 3374.0 0.2X
OpenJDK 64-Bit Server VM 1.8.0_252-8u252-b09-1~18.04-b09 on Linux 4.15.0-1063-aws
Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz
make_interval(): Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
--------------------------------------------------------------------------------------------------------------------------
prepare make_interval() 3634 3684 47 0.3 3634.1 1.0X
make_interval(0, 1, 2, 3, 4, 5, 50.123456) 90 100 12 11.1 90.0 40.4X
make_interval(*, *, 2, 3, 4, 5, 50.123456) 114 119 5 8.8 114.3 31.8X
make_interval(0, 1, *, *, 4, 5, 50.123456) 121 138 21 8.3 120.7 30.1X
make_interval(0, 1, 2, 3, *, *, *) 3615 3621 9 0.3 3614.7 1.0X
make_interval(*, *, *, *, *, *, *) 3638 3657 21 0.3 3637.7 1.0X
[SPARK-29533][SQL][TEST] Benchmark casting strings to intervals ### What changes were proposed in this pull request? Added new benchmark `IntervalBenchmark` to measure performance of interval related functions. In the PR, I added benchmarks for casting strings to interval. In particular, interval strings with `interval` prefix and without it because there is special code for this https://github.com/apache/spark/blob/da576a737c2db01e5ba5ce19ed0e8f900cb5efaf/common/unsafe/src/main/java/org/apache/spark/unsafe/types/CalendarInterval.java#L100-L103 . And also I added benchmarks for different number of units in interval strings, for example 1 unit is `interval 10 years`, 2 units w/o interval is `10 years 5 months`, and etc. ### Why are the changes needed? - To find out current performance issues in casting to intervals - The benchmark can be used while refactoring/re-implementing `CalendarInterval.fromString()` or `CalendarInterval.fromCaseInsensitiveString()`. ### Does this PR introduce any user-facing change? No ### How was this patch tested? By running the benchmark via the command: ```shell SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "sql/test:runMain org.apache.spark.sql.execution.benchmark.IntervalBenchmark" ``` Closes #26189 from MaxGekk/interval-from-string-benchmark. Authored-by: Maxim Gekk <max.gekk@gmail.com> Signed-off-by: HyukjinKwon <gurwls223@apache.org>
2019-10-21 21:47:04 -04:00
Reference in a new issue Copy permalink