Repository: spark Updated Branches: refs/heads/master 4035c98a0 -> 922dfe486
[SPARK-25965][SQL][TEST] Add avro read benchmark Add read benchmark for Avro, which is missing for a period. The benchmark is similar to `DataSourceReadBenchmark` and `OrcReadBenchmark` Manually run benchmark Closes #22966 from gengliangwang/avroReadBenchmark. Lead-authored-by: Gengliang Wang <gengliang.w...@databricks.com> Co-authored-by: Gengliang Wang <ltn...@gmail.com> Co-authored-by: Dongjoon Hyun <dongj...@apache.org> Signed-off-by: Dongjoon Hyun <dongj...@apache.org> Project: http://git-wip-us.apache.org/repos/asf/spark/repo Commit: http://git-wip-us.apache.org/repos/asf/spark/commit/922dfe48 Tree: http://git-wip-us.apache.org/repos/asf/spark/tree/922dfe48 Diff: http://git-wip-us.apache.org/repos/asf/spark/diff/922dfe48 Branch: refs/heads/master Commit: 922dfe4865216987f9a92892b89dc3eaa9610b9b Parents: 4035c98 Author: Gengliang Wang <gengliang.w...@databricks.com> Authored: Wed Nov 14 11:25:11 2018 -0800 Committer: Dongjoon Hyun <dongj...@apache.org> Committed: Wed Nov 14 11:26:26 2018 -0800 ---------------------------------------------------------------------- .../benchmarks/AvroReadBenchmark-results.txt | 122 +++++++++++ .../execution/benchmark/AvroReadBenchmark.scala | 216 +++++++++++++++++++ 2 files changed, 338 insertions(+) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/spark/blob/922dfe48/external/avro/benchmarks/AvroReadBenchmark-results.txt ---------------------------------------------------------------------- diff --git a/external/avro/benchmarks/AvroReadBenchmark-results.txt b/external/avro/benchmarks/AvroReadBenchmark-results.txt new file mode 100644 index 0000000..7900fea --- /dev/null +++ b/external/avro/benchmarks/AvroReadBenchmark-results.txt @@ -0,0 +1,122 @@ +================================================================================================ +SQL Single Numeric Column Scan +================================================================================================ + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +SQL Single TINYINT Column Scan: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum 2774 / 2815 5.7 176.4 1.0X + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +SQL Single SMALLINT Column Scan: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum 2761 / 2777 5.7 175.5 1.0X + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +SQL Single INT Column Scan: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum 2783 / 2870 5.7 176.9 1.0X + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +SQL Single BIGINT Column Scan: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum 3256 / 3266 4.8 207.0 1.0X + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +SQL Single FLOAT Column Scan: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum 2841 / 2867 5.5 180.6 1.0X + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +SQL Single DOUBLE Column Scan: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum 2981 / 2996 5.3 189.5 1.0X + + +================================================================================================ +Int and String Scan +================================================================================================ + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +Int and String Scan: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum of columns 4781 / 4783 2.2 456.0 1.0X + + +================================================================================================ +Partitioned Table Scan +================================================================================================ + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +Partitioned Table: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Data column 3372 / 3386 4.7 214.4 1.0X +Partition column 3035 / 3064 5.2 193.0 1.1X +Both columns 3445 / 3461 4.6 219.1 1.0X + + +================================================================================================ +Repeated String Scan +================================================================================================ + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +Repeated String: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum of string length 3395 / 3401 3.1 323.8 1.0X + + +================================================================================================ +String with Nulls Scan +================================================================================================ + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +String with Nulls Scan (0.0%): Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum of string length 5580 / 5624 1.9 532.2 1.0X + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +String with Nulls Scan (50.0%): Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum of string length 4622 / 4623 2.3 440.8 1.0X + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +String with Nulls Scan (95.0%): Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum of string length 3238 / 3241 3.2 308.8 1.0X + + +================================================================================================ +Single Column Scan From Wide Columns +================================================================================================ + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +Single Column Scan from 100 columns: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum of single column 5472 / 5484 0.2 5218.8 1.0X + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +Single Column Scan from 200 columns: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum of single column 10680 / 10701 0.1 10185.1 1.0X + +OpenJDK 64-Bit Server VM 1.8.0_191-b12 on Linux 3.10.0-862.3.2.el7.x86_64 +Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz +Single Column Scan from 300 columns: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative +------------------------------------------------------------------------------------------------ +Sum of single column 16143 / 16238 0.1 15394.9 1.0X + + http://git-wip-us.apache.org/repos/asf/spark/blob/922dfe48/external/avro/src/test/scala/org/apache/spark/sql/execution/benchmark/AvroReadBenchmark.scala ---------------------------------------------------------------------- diff --git a/external/avro/src/test/scala/org/apache/spark/sql/execution/benchmark/AvroReadBenchmark.scala b/external/avro/src/test/scala/org/apache/spark/sql/execution/benchmark/AvroReadBenchmark.scala new file mode 100644 index 0000000..f2f7d65 --- /dev/null +++ b/external/avro/src/test/scala/org/apache/spark/sql/execution/benchmark/AvroReadBenchmark.scala @@ -0,0 +1,216 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +package org.apache.spark.sql.execution.benchmark + +import java.io.File + +import scala.util.Random + +import org.apache.spark.benchmark.Benchmark +import org.apache.spark.sql.DataFrame +import org.apache.spark.sql.catalyst.plans.SQLHelper +import org.apache.spark.sql.types._ + +/** + * Benchmark to measure Avro read performance. + * {{{ + * To run this benchmark: + * 1. without sbt: bin/spark-submit --class <this class> + * --jars <catalyst test jar>,<core test jar>,<spark-avro jar> <avro test jar> + * 2. build/sbt "avro/test:runMain <this class>" + * 3. generate result: SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "avro/test:runMain <this class>" + * Results will be written to "benchmarks/AvroReadBenchmark-results.txt". + * }}} + */ +object AvroReadBenchmark extends SqlBasedBenchmark with SQLHelper { + def withTempTable(tableNames: String*)(f: => Unit): Unit = { + try f finally tableNames.foreach(spark.catalog.dropTempView) + } + + private def prepareTable(dir: File, df: DataFrame, partition: Option[String] = None): Unit = { + val dirAvro = dir.getCanonicalPath + + if (partition.isDefined) { + df.write.partitionBy(partition.get).format("avro").save(dirAvro) + } else { + df.write.format("avro").save(dirAvro) + } + + spark.read.format("avro").load(dirAvro).createOrReplaceTempView("avroTable") + } + + def numericScanBenchmark(values: Int, dataType: DataType): Unit = { + val benchmark = + new Benchmark(s"SQL Single ${dataType.sql} Column Scan", values, output = output) + + withTempPath { dir => + withTempTable("t1", "avroTable") { + import spark.implicits._ + spark.range(values).map(_ => Random.nextLong).createOrReplaceTempView("t1") + + prepareTable(dir, spark.sql(s"SELECT CAST(value as ${dataType.sql}) id FROM t1")) + + benchmark.addCase("Sum") { _ => + spark.sql("SELECT sum(id) FROM avroTable").collect() + } + + benchmark.run() + } + } + } + + def intStringScanBenchmark(values: Int): Unit = { + val benchmark = new Benchmark("Int and String Scan", values, output = output) + + withTempPath { dir => + withTempTable("t1", "avroTable") { + import spark.implicits._ + spark.range(values).map(_ => Random.nextLong).createOrReplaceTempView("t1") + + prepareTable( + dir, + spark.sql("SELECT CAST(value AS INT) AS c1, CAST(value as STRING) AS c2 FROM t1")) + + benchmark.addCase("Sum of columns") { _ => + spark.sql("SELECT sum(c1), sum(length(c2)) FROM avroTable").collect() + } + + benchmark.run() + } + } + } + + def partitionTableScanBenchmark(values: Int): Unit = { + val benchmark = new Benchmark("Partitioned Table", values, output = output) + + withTempPath { dir => + withTempTable("t1", "avroTable") { + import spark.implicits._ + spark.range(values).map(_ => Random.nextLong).createOrReplaceTempView("t1") + + prepareTable(dir, spark.sql("SELECT value % 2 AS p, value AS id FROM t1"), Some("p")) + + benchmark.addCase("Data column") { _ => + spark.sql("SELECT sum(id) FROM avroTable").collect() + } + + benchmark.addCase("Partition column") { _ => + spark.sql("SELECT sum(p) FROM avroTable").collect() + } + + benchmark.addCase("Both columns") { _ => + spark.sql("SELECT sum(p), sum(id) FROM avroTable").collect() + } + + benchmark.run() + } + } + } + + def repeatedStringScanBenchmark(values: Int): Unit = { + val benchmark = new Benchmark("Repeated String", values, output = output) + + withTempPath { dir => + withTempTable("t1", "avroTable") { + spark.range(values).createOrReplaceTempView("t1") + + prepareTable(dir, spark.sql("SELECT CAST((id % 200) + 10000 as STRING) AS c1 FROM t1")) + + benchmark.addCase("Sum of string length") { _ => + spark.sql("SELECT sum(length(c1)) FROM avroTable").collect() + } + + benchmark.run() + } + } + } + + def stringWithNullsScanBenchmark(values: Int, fractionOfNulls: Double): Unit = { + withTempPath { dir => + withTempTable("t1", "avroTable") { + spark.range(values).createOrReplaceTempView("t1") + + prepareTable( + dir, + spark.sql( + s"SELECT IF(RAND(1) < $fractionOfNulls, NULL, CAST(id as STRING)) AS c1, " + + s"IF(RAND(2) < $fractionOfNulls, NULL, CAST(id as STRING)) AS c2 FROM t1")) + + val percentageOfNulls = fractionOfNulls * 100 + val benchmark = + new Benchmark(s"String with Nulls Scan ($percentageOfNulls%)", values, output = output) + + benchmark.addCase("Sum of string length") { _ => + spark.sql("SELECT SUM(LENGTH(c2)) FROM avroTable " + + "WHERE c1 IS NOT NULL AND c2 IS NOT NULL").collect() + } + + benchmark.run() + } + } + } + + def columnsBenchmark(values: Int, width: Int): Unit = { + val benchmark = + new Benchmark(s"Single Column Scan from $width columns", values, output = output) + + withTempPath { dir => + withTempTable("t1", "avroTable") { + import spark.implicits._ + val middle = width / 2 + val selectExpr = (1 to width).map(i => s"value as c$i") + spark.range(values).map(_ => Random.nextLong).toDF() + .selectExpr(selectExpr: _*).createOrReplaceTempView("t1") + + prepareTable(dir, spark.sql("SELECT * FROM t1")) + + benchmark.addCase("Sum of single column") { _ => + spark.sql(s"SELECT sum(c$middle) FROM avroTable").collect() + } + + benchmark.run() + } + } + } + + override def runBenchmarkSuite(mainArgs: Array[String]): Unit = { + runBenchmark("SQL Single Numeric Column Scan") { + Seq(ByteType, ShortType, IntegerType, LongType, FloatType, DoubleType).foreach { dataType => + numericScanBenchmark(1024 * 1024 * 15, dataType) + } + } + runBenchmark("Int and String Scan") { + intStringScanBenchmark(1024 * 1024 * 10) + } + runBenchmark("Partitioned Table Scan") { + partitionTableScanBenchmark(1024 * 1024 * 15) + } + runBenchmark("Repeated String Scan") { + repeatedStringScanBenchmark(1024 * 1024 * 10) + } + runBenchmark("String with Nulls Scan") { + for (fractionOfNulls <- List(0.0, 0.50, 0.95)) { + stringWithNullsScanBenchmark(1024 * 1024 * 10, fractionOfNulls) + } + } + runBenchmark("Single Column Scan From Wide Columns") { + columnsBenchmark(1024 * 1024 * 1, 100) + columnsBenchmark(1024 * 1024 * 1, 200) + columnsBenchmark(1024 * 1024 * 1, 300) + } + } +} --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@spark.apache.org For additional commands, e-mail: commits-h...@spark.apache.org
