From 8d35a6f68d6d733212674491cbf31bed73fada0f Mon Sep 17 00:00:00 2001 From: Davies Liu Date: Mon, 22 Aug 2016 16:16:03 +0800 Subject: [SPARK-17115][SQL] decrease the threshold when split expressions ## What changes were proposed in this pull request? In 2.0, we change the threshold of splitting expressions from 16K to 64K, which cause very bad performance on wide table, because the generated method can't be JIT compiled by default (above the limit of 8K bytecode). This PR will decrease it to 1K, based on the benchmark results for a wide table with 400 columns of LongType. It also fix a bug around splitting expression in whole-stage codegen (it should not split them). ## How was this patch tested? Added benchmark suite. Author: Davies Liu Closes #14692 from davies/split_exprs. --- .../execution/aggregate/HashAggregateExec.scala | 2 - .../execution/benchmark/BenchmarkWideTable.scala | 53 ++++++++++++++++++++++ 2 files changed, 53 insertions(+), 2 deletions(-) create mode 100644 sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/BenchmarkWideTable.scala (limited to 'sql/core/src') diff --git a/sql/core/src/main/scala/org/apache/spark/sql/execution/aggregate/HashAggregateExec.scala b/sql/core/src/main/scala/org/apache/spark/sql/execution/aggregate/HashAggregateExec.scala index cfc47aba88..bd7efa606e 100644 --- a/sql/core/src/main/scala/org/apache/spark/sql/execution/aggregate/HashAggregateExec.scala +++ b/sql/core/src/main/scala/org/apache/spark/sql/execution/aggregate/HashAggregateExec.scala @@ -603,8 +603,6 @@ case class HashAggregateExec( // create grouping key ctx.currentVars = input - // make sure that the generated code will not be splitted as multiple functions - ctx.INPUT_ROW = null val unsafeRowKeyCode = GenerateUnsafeProjection.createCode( ctx, groupingExpressions.map(e => BindReferences.bindReference[Expression](e, child.output))) val vectorizedRowKeys = ctx.generateExpressions( diff --git a/sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/BenchmarkWideTable.scala b/sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/BenchmarkWideTable.scala new file mode 100644 index 0000000000..9dcaca0ca9 --- /dev/null +++ b/sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/BenchmarkWideTable.scala @@ -0,0 +1,53 @@ +/* + * 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 org.apache.spark.util.Benchmark + + +/** + * Benchmark to measure performance for wide table. + * To run this: + * build/sbt "sql/test-only *benchmark.BenchmarkWideTable" + * + * Benchmarks in this file are skipped in normal builds. + */ +class BenchmarkWideTable extends BenchmarkBase { + + ignore("project on wide table") { + val N = 1 << 20 + val df = sparkSession.range(N) + val columns = (0 until 400).map{ i => s"id as id$i"} + val benchmark = new Benchmark("projection on wide table", N) + benchmark.addCase("wide table", numIters = 5) { iter => + df.selectExpr(columns : _*).queryExecution.toRdd.count() + } + benchmark.run() + + /** + * Here are some numbers with different split threshold: + * + * Split threshold methods Rate(M/s) Per Row(ns) + * 10 400 0.4 2279 + * 100 200 0.6 1554 + * 1k 37 0.9 1116 + * 8k 5 0.5 2025 + * 64k 1 0.0 21649 + */ + } +} -- cgit v1.2.3