Github user cloud-fan commented on a diff in the pull request:
https://github.com/apache/spark/pull/13680#discussion_r69850526
--- Diff:
sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/UnsafeArrayDataBenchmark.scala
---
@@ -0,0 +1,251 @@
+/*
+ * 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 scala.util.Random
+
+import org.apache.spark.sql.catalyst.encoders.ExpressionEncoder
+import org.apache.spark.sql.catalyst.expressions.{UnsafeArrayData,
UnsafeRow}
+import org.apache.spark.sql.catalyst.expressions.codegen.{BufferHolder,
UnsafeArrayWriter}
+import org.apache.spark.util.Benchmark
+
+/**
+ * Benchmark [[UnsafeArrayDataBenchmark]] for UnsafeArrayData
+ * To run this:
+ * 1. replace ignore(...) with test(...)
+ * 2. build/sbt "sql/test-only *benchmark.UnsafeArrayDataBenchmark"
+ *
+ * Benchmarks in this file are skipped in normal builds.
+ */
+class UnsafeArrayDataBenchmark extends BenchmarkBase {
+
+ def calculateHeaderPortionInBytes(count: Int) : Int = {
+ // Use this assignment for SPARK-15962
+ // val size = 4 + 4 * count
+ val size = UnsafeArrayData.calculateHeaderPortionInBytes(count)
+ size
+ }
+
+ def readUnsafeArray(iters: Int): Unit = {
+ val count = 1024 * 1024 * 16
+ val rand = new Random(42)
+
+ var intResult: Int = 0
+ val intBuffer = Array.fill[Int](count) { rand.nextInt }
+ val intEncoder = ExpressionEncoder[Array[Int]].resolveAndBind()
+ val intInternalRow = intEncoder.toRow(intBuffer)
+ val intUnsafeArray = intInternalRow.getArray(0)
+ val readIntArray = { i: Int =>
+ var n = 0
+ while (n < iters) {
+ val len = intUnsafeArray.numElements
+ var sum = 0.toInt
+ var i = 0
+ while (i < len) {
+ sum += intUnsafeArray.getInt(i)
+ i += 1
+ }
+ intResult = sum
+ n += 1
+ }
+ }
+
+ var doubleResult: Double = 0
+ val doubleBuffer = Array.fill[Double](count) { rand.nextDouble }
+ val doubleEncoder = ExpressionEncoder[Array[Double]].resolveAndBind()
+ val doubleInternalRow = doubleEncoder.toRow(doubleBuffer)
+ val doubleUnsafeArray = doubleInternalRow.getArray(0)
+ val readDoubleArray = { i: Int =>
+ var n = 0
+ while (n < iters) {
+ val len = doubleUnsafeArray.numElements
+ var sum = 0.toDouble
+ var i = 0
+ while (i < len) {
+ sum += doubleUnsafeArray.getDouble(i)
+ i += 1
+ }
+ doubleResult = sum
+ n += 1
+ }
+ }
+
+ val benchmark = new Benchmark("Read UnsafeArrayData", count * iters)
+ benchmark.addCase("Int")(readIntArray)
+ benchmark.addCase("Double")(readDoubleArray)
+ benchmark.run
+ /*
+ Java HotSpot(TM) 64-Bit Server VM 1.8.0_92-b14 on Mac OS X 10.10.4
+ Intel(R) Core(TM) i5-5257U CPU @ 2.70GHz
+
+ Read UnsafeArrayData: Best/Avg Time(ms)
Rate(M/s) Per Row(ns) Relative
+
------------------------------------------------------------------------------------------------
+ Int 279 / 294 600.4
1.7 1.0X
+ Double 296 / 303 567.0
1.8 0.9X
+ */
+ }
+
+ def writeUnsafeArray(iters: Int): Unit = {
+ val count = 1024 * 1024 * 16
+
+ val intUnsafeRow = new UnsafeRow(1)
+ val intUnsafeArrayWriter = new UnsafeArrayWriter
--- End diff --
have you seen my comment here?
https://github.com/apache/spark/pull/13680/files#r69392823
testing the array writer is so low level and peoples are more interested in
writing the whole array. If you take a look at what `encoder.toRow` does, it
generates a project to write the given array into an unsafe row. Although it
has some overhead for the row stuff, it's still a good example of writing array.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
