From 21f9c32dc685d7d3a3d637ac6b7ea0c66f2d8681 Mon Sep 17 00:00:00 2001
From: seancxmao <seancxmao@gmail.com>
Date: Thu, 20 Sep 2018 23:32:17 +0800
Subject: [PATCH] [SPARK-25489] Refactor UDTSerializationBenchmark

---
 .../UDTSerializationBenchmark-results.txt     | 13 ++++
 .../linalg/UDTSerializationBenchmark.scala    | 70 ++++++++++---------
 2 files changed, 49 insertions(+), 34 deletions(-)
 create mode 100644 mllib/benchmarks/UDTSerializationBenchmark-results.txt

diff --git a/mllib/benchmarks/UDTSerializationBenchmark-results.txt b/mllib/benchmarks/UDTSerializationBenchmark-results.txt
new file mode 100644
index 0000000000000..169f4c60c748e
--- /dev/null
+++ b/mllib/benchmarks/UDTSerializationBenchmark-results.txt
@@ -0,0 +1,13 @@
+================================================================================================
+VectorUDT de/serialization
+================================================================================================
+
+Java HotSpot(TM) 64-Bit Server VM 1.8.0_131-b11 on Mac OS X 10.13.6
+Intel(R) Core(TM) i7-6820HQ CPU @ 2.70GHz
+
+VectorUDT de/serialization:              Best/Avg Time(ms)    Rate(M/s)   Per Row(ns)   Relative
+------------------------------------------------------------------------------------------------
+serialize                                      144 /  206          0.0      143979.7       1.0X
+deserialize                                    114 /  135          0.0      113802.6       1.3X
+
+
diff --git a/mllib/src/test/scala/org/apache/spark/mllib/linalg/UDTSerializationBenchmark.scala b/mllib/src/test/scala/org/apache/spark/mllib/linalg/UDTSerializationBenchmark.scala
index e2976e1ab022b..1a2216ea070c4 100644
--- a/mllib/src/test/scala/org/apache/spark/mllib/linalg/UDTSerializationBenchmark.scala
+++ b/mllib/src/test/scala/org/apache/spark/mllib/linalg/UDTSerializationBenchmark.scala
@@ -17,53 +17,55 @@
 
 package org.apache.spark.mllib.linalg
 
-import org.apache.spark.benchmark.Benchmark
+import org.apache.spark.benchmark.{Benchmark, BenchmarkBase}
 import org.apache.spark.sql.catalyst.encoders.ExpressionEncoder
 
 /**
  * Serialization benchmark for VectorUDT.
+ * To run this benchmark:
+ * {{{
+ * 1. without sbt: bin/spark-submit --class <this class> <spark mllib test jar>
+ * 2. build/sbt "mllib/test:runMain <this class>"
+ * 3. generate result: SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "mllib/test:runMain <this class>"
+ *    Results will be written to "benchmarks/UDTSerializationBenchmark-results.txt".
+ * }}}
  */
-object UDTSerializationBenchmark {
+object UDTSerializationBenchmark extends BenchmarkBase {
 
-  def main(args: Array[String]): Unit = {
-    val iters = 1e2.toInt
-    val numRows = 1e3.toInt
+  override def benchmark(): Unit = {
 
-    val encoder = ExpressionEncoder[Vector].resolveAndBind()
+    runBenchmark("VectorUDT de/serialization") {
+      val iters = 1e2.toInt
+      val numRows = 1e3.toInt
 
-    val vectors = (1 to numRows).map { i =>
-      Vectors.dense(Array.fill(1e5.toInt)(1.0 * i))
-    }.toArray
-    val rows = vectors.map(encoder.toRow)
+      val encoder = ExpressionEncoder[Vector].resolveAndBind()
 
-    val benchmark = new Benchmark("VectorUDT de/serialization", numRows, iters)
+      val vectors = (1 to numRows).map { i =>
+        Vectors.dense(Array.fill(1e5.toInt)(1.0 * i))
+      }.toArray
+      val rows = vectors.map(encoder.toRow)
 
-    benchmark.addCase("serialize") { _ =>
-      var sum = 0
-      var i = 0
-      while (i < numRows) {
-        sum += encoder.toRow(vectors(i)).numFields
-        i += 1
+      val benchmark = new Benchmark("VectorUDT de/serialization", numRows, iters, output = output)
+
+      benchmark.addCase("serialize") { _ =>
+        var sum = 0
+        var i = 0
+        while (i < numRows) {
+          sum += encoder.toRow(vectors(i)).numFields
+          i += 1
+        }
       }
-    }
 
-    benchmark.addCase("deserialize") { _ =>
-      var sum = 0
-      var i = 0
-      while (i < numRows) {
-        sum += encoder.fromRow(rows(i)).numActives
-        i += 1
+      benchmark.addCase("deserialize") { _ =>
+        var sum = 0
+        var i = 0
+        while (i < numRows) {
+          sum += encoder.fromRow(rows(i)).numActives
+          i += 1
+        }
       }
-    }
 
-    /*
-    OpenJDK 64-Bit Server VM 1.8.0_91-b14 on Linux 4.4.11-200.fc22.x86_64
-    Intel Xeon E3-12xx v2 (Ivy Bridge)
-    VectorUDT de/serialization:              Best/Avg Time(ms)    Rate(M/s)   Per Row(ns)   Relative
-    ------------------------------------------------------------------------------------------------
-    serialize                                      265 /  318          0.0      265138.5       1.0X
-    deserialize                                    155 /  197          0.0      154611.4       1.7X
-    */
-    benchmark.run()
+      benchmark.run()
+    }
   }
 }