On Thu, 20 Jun 2024 14:32:25 GMT, Shaojin Wen <d...@openjdk.org> wrote:
>> @cl4es @wenshao I think we should probably work on `putChar`, or at least
>> figure out what blocks `MergeStore` for `putChar`. Can someone produce a
>> simple stand-alone `.java` file that uses `putChar`, so that I can
>> investigate why `MergeStore` does not work for it?
>>
>> Otherwise, I agree with @cl4es : the code here may be ok for now, but we
>> would have to revert it again later, should `MergeStore` eventually do the
>> trick.
>
> @eme64
>
> simple stand-alone java
>
> import jdk.internal.misc.Unsafe;
>
> public class PutCharTest {
> static final Unsafe UNSAFE = Unsafe.getUnsafe();
>
> static void putCharsAt(byte[] val, int index, int c1, int c2, int c3, int
> c4) {
> putChar(val, index , (char)(c1));
> putChar(val, index + 1, (char)(c2));
> putChar(val, index + 2, (char)(c3));
> putChar(val, index + 3, (char)(c4));
> }
>
> static void putChar(byte[] bytes, int index, int c) {
> UNSAFE.putChar(bytes, Unsafe.ARRAY_BYTE_BASE_OFFSET + (index << 1),
> (char)(c));
> }
>
> static void putChar0(byte[] bytes, int index, int c) {
> UNSAFE.putByte(bytes, Unsafe.ARRAY_BYTE_BASE_OFFSET + (index << 1),
> (byte)(c));
> UNSAFE.putByte(bytes, Unsafe.ARRAY_BYTE_BASE_OFFSET + (index << 1) +
> 1, (byte)(c << 8));
> }
>
> static void putNull(byte[] bytes, int index) {
> putCharsAt(bytes, index, 'n', 'u', 'l', 'l');
> }
>
> public static void main(String[] args) {
> for (int i = 0; i < 5; i++) {
> testNull();
> }
>
> System.out.println("done");
> }
>
> private static void testNull() {
> byte[] bytes = new byte[8192];
>
> for (int i = 0; i < 1000; i++) {
> int index = 0;
> for (int j = 0; j < 1024; j++) {
> putNull(bytes, index);
> index += 4;
> }
> }
> }
> }
@wenshao @cl4es I think this use-case is quite a valid one, and deserves an
extension. I filed:
[JDK-8335113](https://bugs.openjdk.org/browse/JDK-8335113): C2 MergeStores:
allow merging of putChar and other larger stores on smaller array elements
-------------
PR Comment: https://git.openjdk.org/jdk/pull/19626#issuecomment-2189523984
