1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236
|
/*
* Copyright (c) 2020, 2024, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* This code is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License version 2 only, as
* published by the Free Software Foundation.
*
* This code is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* version 2 for more details (a copy is included in the LICENSE file that
* accompanied this code).
*
* You should have received a copy of the GNU General Public License version
* 2 along with this work; if not, write to the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
*
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
* or visit www.oracle.com if you need additional information or have any
* questions.
*/
/**
* @test
* @bug 8257531 8310190
* @summary Test vectorization for Buffer operations.
* @library /test/lib /
* @requires vm.compiler2.enabled
* @run driver compiler.vectorization.TestBufferVectorization
*/
package compiler.vectorization;
import compiler.lib.ir_framework.*;
import java.lang.invoke.MethodHandles;
import java.lang.invoke.VarHandle;
import java.nio.ByteBuffer;
import java.nio.ByteOrder;
import java.nio.IntBuffer;
public class TestBufferVectorization {
final static int N = 1024*16;
static int offset = 0;
final static VarHandle VH_arr_view = MethodHandles.byteArrayViewVarHandle(int[].class, ByteOrder.nativeOrder()).withInvokeExactBehavior();
public static void main(String[] args) {
TestFramework.run();
}
@Run(test = "testArray")
public static void runArray() {
int[] array = new int[N];
for (int k = 0; k < array.length; k++) {
array[k] = k;
}
testArray(array);
for(int k = 0; k < array.length; k++) {
if (array[k] != (k + 1)) {
throw new RuntimeException(" Invalid result: array[" + k + "]: " + array[k] + " != " + (k + 1));
}
}
}
@Test
@IR(counts = {IRNode.REPLICATE_I, ">0",
IRNode.LOAD_VECTOR_I, ">0",
IRNode.ADD_VI, ">0",
IRNode.STORE_VECTOR, ">0"},
applyIfCPUFeatureOr = {"sse4.1", "true", "asimd", "true", "rvv", "true"})
public static void testArray(int[] array) {
for(int k = 0; k < array.length; k++) {
array[k] += 1;
}
}
@Run(test = "testArrayOffset")
public static void runArrayOffset() {
// Moving offset between 0..255
offset = (offset + 1) % 256;
int[] array = new int[N];
for (int k = 0; k < array.length; k++) {
array[k] = k;
}
testArrayOffset(array, offset);
int l = array.length - offset;
for(int k = 0; k < offset; k++) {
if (array[k] != k) {
throw new RuntimeException(" Invalid result: arrayOffset[" + k + "]: " + array[k] + " != " + (k + 1));
}
}
for(int k = offset; k < array.length; k++) {
if (array[k] != (k + 1)) {
throw new RuntimeException(" Invalid result: arrayOffset[" + k + "]: " + array[k] + " != " + k);
}
}
}
@Test
@IR(counts = {IRNode.REPLICATE_I, ">0",
IRNode.LOAD_VECTOR_I, ">0",
IRNode.ADD_VI, ">0",
IRNode.STORE_VECTOR, ">0"},
applyIfCPUFeatureOr = {"sse4.1", "true", "asimd", "true", "rvv", "true"})
public static void testArrayOffset(int[] array, int offset) {
int l = array.length - offset;
for(int k = 0; k < l; k++) {
array[k + offset] += 1;
}
}
@Run(test = "testBuffer")
public static void runBuffer() {
IntBuffer buffer = IntBuffer.allocate(N);
initBuffer(buffer);
testBuffer(buffer);
verifyBuffer(buffer);
}
@Test
@IR(counts = {IRNode.REPLICATE_I, ">0",
IRNode.LOAD_VECTOR_I, ">0",
IRNode.ADD_VI, ">0",
IRNode.STORE_VECTOR, ">0"},
applyIfCPUFeatureOr = {"sse4.1", "true", "asimd", "true", "rvv", "true"})
public static void testBuffer(IntBuffer buffer) {
for (int k = 0; k < buffer.limit(); k++) {
buffer.put(k, buffer.get(k) + 1);
}
}
@Run(test = "testBufferHeap")
public static void runBufferHeap() {
IntBuffer buffer = ByteBuffer.allocate(N * Integer.BYTES).order(ByteOrder.nativeOrder()).asIntBuffer();
initBuffer(buffer);
testBufferHeap(buffer);
verifyBuffer(buffer);
}
@Test
@IR(counts = {IRNode.REPLICATE_I, IRNode.VECTOR_SIZE_ANY, ">0",
IRNode.LOAD_VECTOR_I, IRNode.VECTOR_SIZE_ANY, ">0",
IRNode.ADD_VI, IRNode.VECTOR_SIZE_ANY, ">0",
IRNode.STORE_VECTOR, ">0"},
applyIfCPUFeatureOr = {"sse4.1", "true", "asimd", "true", "rvv", "true"},
applyIf = {"AlignVector", "false"},
applyIfPlatform = {"64-bit", "true"})
// VECTOR_SIZE_ANY: Unrolling does not always seem to go far enough to reach maximum vector size.
// This looks like a BUG.
// AlignVector: Buffer get/put have an invariant that is in bytes (LoadL in ByteBufferAsIntBufferL::byteOffset).
// This makes sense: we are accessing a byte buffer. But to be able to align the 4 byte ints,
// we would require to know that the invariant is a multiple of 4. Without that, we cannot
// guarantee alignment by adjusting the limit of the pre-loop with a stride of 4 bytes.
// 64-bit: bufferHeap uses Long type for memory accesses which are not vectorized in 32-bit VM
public static void testBufferHeap(IntBuffer buffer) {
for (int k = 0; k < buffer.limit(); k++) {
buffer.put(k, buffer.get(k) + 1);
}
}
@Run(test = "testBufferDirect")
public static void runBufferDirect() {
IntBuffer buffer = ByteBuffer.allocateDirect(N * Integer.BYTES).order(ByteOrder.nativeOrder()).asIntBuffer();
initBuffer(buffer);
testBufferDirect(buffer);
verifyBuffer(buffer);
}
@Test
// bufferDirect uses Unsafe memory accesses which are not vectorized currently
// We find a CastX2P in pointer analysis (VPointer)
public static void testBufferDirect(IntBuffer buffer) {
for (int k = 0; k < buffer.limit(); k++) {
buffer.put(k, buffer.get(k) + 1);
}
}
public static void initBuffer(IntBuffer buffer) {
for (int k = 0; k < buffer.limit(); k++) {
buffer.put(k, k);
}
}
public static void verifyBuffer(IntBuffer buffer) {
for(int k = 0; k < buffer.limit(); k++) {
if (buffer.get(k) != (k + 1)) {
throw new RuntimeException(" Invalid result: buffer.get(" + k + "): " + buffer.get(k) + " != " + (k + 1));
}
}
}
@Run(test = "testArrayView")
public static void runArrayView() {
byte[] b_arr = new byte[N * Integer.BYTES];
for (int k = 0; k < N; k++) {
VH_arr_view.set(b_arr, k, k);
}
// Save initial INT values
int[] i_arr = new int[N];
for (int k = 0; k < i_arr.length; k++) {
i_arr[k] = (int) VH_arr_view.get(b_arr, k * Integer.BYTES);
}
testArrayView(b_arr);
for (int k = 0; k < i_arr.length; k++) {
int v = (int) VH_arr_view.get(b_arr, k * Integer.BYTES);
if (v != (i_arr[k] + 1)) {
throw new RuntimeException(" Invalid result: VH_arr_view.get(b_arr, " + (k * Integer.BYTES) + "): " + v + " != " + (i_arr[k] + 1));
}
}
}
@Test
@IR(counts = {IRNode.REPLICATE_I, ">0",
IRNode.LOAD_VECTOR_I, ">0",
IRNode.ADD_VI, ">0",
IRNode.STORE_VECTOR, ">0"},
applyIfCPUFeatureOr = {"sse4.1", "true", "asimd", "true", "rvv", "true"})
public static void testArrayView(byte[] b_arr) {
for (int k = 0; k < b_arr.length; k += 4) {
int v = (int) VH_arr_view.get(b_arr, k);
VH_arr_view.set(b_arr, k, v + 1);
}
}
}
|