Add samples to measure vector multiply-add performance

Add the test cases as follows:
1. Vector sum of multiply-add for int
2. Vector sum of multiply-add and vector multiply-add for short

SLP support for different data types can also be measured by comparing
short cases with int cases.

Change-Id: Ia57f80c3b21edd0fcc78866fac8b848b4a2d1b4f

2 files changed, 220 insertions, 0 deletions

diff --git a/src/main/java/org/linaro/benchmarks/TestSIMDMlaInt.java b/src/main/java/org/linaro/benchmarks/TestSIMDMlaInt.java
new file mode 100644
index 0000000..ae1bc6f
--- /dev/null
+++ b/src/main/java/org/linaro/benchmarks/TestSIMDMlaInt.java
@@ -0,0 +1,88 @@
+package org.linaro.benchmarks;
+
+import org.openjdk.jmh.annotations.*;
+import java.util.concurrent.TimeUnit;
+
+@BenchmarkMode(Mode.AverageTime)
+@OutputTimeUnit(TimeUnit.MICROSECONDS)
+@State(Scope.Benchmark)
+
+public class TestSIMDMlaInt {
+  static final int LENGTH = 256 * 1024;
+  static int [] a = new int[LENGTH];
+  static int [] b = new int[LENGTH];
+  static int [] c = new int[LENGTH];
+  static int [] d = new int[LENGTH];
+
+  public static void TestSIMDMlaInit() {
+    for (int i = 0; i < LENGTH; i++) {
+       a[i] = 3;
+       b[i] = 2;
+       c[i] = 1;
+    }
+  }
+
+  public static int vectSumOfMulAdd1(
+          int[] a,
+          int[] b,
+          int[] c,
+          int[] d) {
+    int total = 0;
+    for (int i = 0; i < LENGTH; i++) {
+      d[i] = (int)(a[i] * b[i] + c[i]);
+      total += d[i];
+    }
+    return total;
+  }
+
+  public static int vectSumOfMulAdd2(
+          int[] a,
+          int[] b,
+          int[] c,
+          int[] d) {
+    int total = 0;
+    for (int i = 0; i < LENGTH; i++) {
+      d[i] = (int)(a[i] * b[i] + a[i] * c[i]);
+      total += d[i];
+    }
+    return total;
+  }
+
+  public static int vectSumOfMulAdd3(
+          int[] a,
+          int[] b,
+          int[] c,
+          int[] d) {
+    int total = 0;
+    for (int i = 0; i < LENGTH; i++) {
+      d[i] = (int)((a[i] * b[i]) + (a[i] * c[i]) + (b[i] * c[i]));
+      total += d[i];
+    }
+    return total;
+  }
+
+  @Setup
+  public void setup()
+  {
+    TestSIMDMlaInit();
+  }
+
+  @Benchmark
+  public void testVectSumOfMulAdd1() {
+    int sum = 0;
+    sum = vectSumOfMulAdd1(a, b, c, d);
+  }
+
+  @Benchmark
+  public void testVectSumOfMulAdd2() {
+    int sum = 0;
+    sum = vectSumOfMulAdd2(a, b, c, d);
+  }
+
+  @Benchmark
+  public void testVectSumOfMulAdd3() {
+    int sum = 0;
+    sum = vectSumOfMulAdd3(a, b, c, d);
+  }
+
+}
diff --git a/src/main/java/org/linaro/benchmarks/TestSIMDMlaShort.java b/src/main/java/org/linaro/benchmarks/TestSIMDMlaShort.java
new file mode 100644
index 0000000..17cf3ce
--- /dev/null
+++ b/src/main/java/org/linaro/benchmarks/TestSIMDMlaShort.java
@@ -0,0 +1,132 @@
+package org.linaro.benchmarks;
+
+import org.openjdk.jmh.annotations.*;
+import java.util.concurrent.TimeUnit;
+
+@BenchmarkMode(Mode.AverageTime)
+@OutputTimeUnit(TimeUnit.MICROSECONDS)
+@State(Scope.Benchmark)
+
+public class TestSIMDMlaShort {
+  static final int LENGTH = 256 * 1024;
+  static short [] a = new short[LENGTH];
+  static short [] b = new short[LENGTH];
+  static short [] c = new short[LENGTH];
+  static short [] d = new short[LENGTH];
+
+  public static void TestSIMDMlaInit() {
+    for (int i = 0; i < LENGTH; i++) {
+       a[i] = 3;
+       b[i] = 2;
+       c[i] = 1;
+    }
+  }
+
+  public static int vectSumOfMulAdd1(
+          short[] a,
+          short[] b,
+          short[] c,
+          short[] d) {
+    int total = 0;
+    for (int i = 0; i < LENGTH; i++) {
+      d[i] = (short)(a[i] * b[i] + c[i]);
+      total += d[i];
+    }
+    return total;
+  }
+
+  public static int vectSumOfMulAdd2(
+          short[] a,
+          short[] b,
+          short[] c,
+          short[] d) {
+    int total = 0;
+    for (int i = 0; i < LENGTH; i++) {
+      d[i] = (short)(a[i] * b[i] + a[i] * c[i]);
+      total += d[i];
+    }
+    return total;
+  }
+
+  public static int vectSumOfMulAdd3(
+          short[] a,
+          short[] b,
+          short[] c,
+          short[] d) {
+    int total = 0;
+    for (int i = 0; i < LENGTH; i++) {
+      d[i] = (short)((a[i] * b[i]) + (a[i] * c[i]) + (b[i] * c[i]));
+      total += d[i];
+    }
+    return total;
+  }
+
+  public static void vectMulAdd1(
+          short[] a,
+          short[] b,
+          short[] c,
+          short[] d) {
+    for (int i = 0; i < LENGTH; i++) {
+      d[i] = (short)(a[i] * b[i] + c[i]);
+    }
+  }
+
+  public static void vectMulAdd2(
+          short[] a,
+          short[] b,
+          short[] c,
+          short[] d) {
+    for (int i = 0; i < LENGTH; i++) {
+      d[i] = (short)(a[i] * b[i] + a[i] * c[i]);
+    }
+  }
+
+  public static void vectMulAdd3(
+          short[] a,
+          short[] b,
+          short[] c,
+          short[] d) {
+    for (int i = 0; i < LENGTH; i++) {
+      d[i] = (short)((a[i] * b[i]) + (a[i] * c[i]) + (b[i] * c[i]));
+    }
+  }
+
+  @Setup
+  public void setup()
+  {
+    TestSIMDMlaInit();
+  }
+
+  @Benchmark
+  public void testVectSumOfMulAdd1() {
+    int sum = 0;
+    sum = vectSumOfMulAdd1(a, b, c, d);
+  }
+
+  @Benchmark
+  public void testVectSumOfMulAdd2() {
+    int sum = 0;
+    sum = vectSumOfMulAdd2(a, b, c, d);
+  }
+
+  @Benchmark
+  public void testVectSumOfMulAdd3() {
+    int sum = 0;
+    sum = vectSumOfMulAdd3(a, b, c, d);
+  }
+  @Benchmark
+  public void testVectMulAdd1() {
+    vectMulAdd1(a, b, c, d);
+  }
+
+  @Benchmark
+  public void testVectMulAdd2() {
+    vectMulAdd2(a, b, c, d);
+  }
+
+  @Benchmark
+  public void testVectMulAdd3() {
+    vectMulAdd3(a, b, c, d);
+  }
+
+}