RFR: 8257436: [aarch64] Regressions in ArrayCopyUnalignedDst.testByte/testChar for 65-78 bytes when UseSIMDForMemoryOps is on [v3]

[Volker Simonis]

On Wed, 2 Dec 2020 15:02:09 GMT, Evgeny Astigeevich <github.com+42899633+eastig at openjdk.org> wrote:

>> Copying data in range 65 - 80 bytes with
>> ldpq
>> ldpq 
>> ldp
>> stpq
>> stpq
>> stp
>> is faster than with
>> ldpq
>> ldpq
>> ldpq
>> stpq
>> stpq
>> stpq
>> when pointers involved are not properly aligned. This mostly affects byte and char arraycopy. 
>> This patch fixes regressions of ArrayCopyUnalignedDst.testByte/testChar when UseSIMDForMemoryOps is on.
>> See the attached results for performance data
>> [test_byte_results.txt](https://github.com/openjdk/jdk/files/5629676/test_byte_results.txt)
>> [test_char_results.txt](https://github.com/openjdk/jdk/files/5629673/test_char_results.txt)
>
> Evgeny Astigeevich has refreshed the contents of this pull request, and previous commits have been removed. The incremental views will show differences compared to the previous content of the PR.

Hi Evegeny,

thanks for fixing this. Your change looks good to me. Thanks especially for measuring all the different cases.

I only have two minor request regarding the comments. Once fixed I'll sponsor this change.

Best regards,
Volker

src/hotspot/cpu/aarch64/stubGenerator_aarch64.cpp line 1158:

> 1156:       __ ldpq(v2, v3, Address(s, 32));
> 1157:       // Unaligned pointers can be an issue for copying.
> 1158:       // This have more chances to happen when granularity of data is

I think this should read "These have more.." or "This has more.."

src/hotspot/cpu/aarch64/stubGenerator_aarch64.cpp line 1100:

> 1098:     const Register t0 = r3, t1 = r4;
> 1099: 
> 1100:     // <= 96 bytes do inline. Direction doesn't matter because we always

Can you please add both cases into the comment. Something like:

// <= 80 (or 96 for SIMD) do inline...

-------------

Changes requested by simonis (Reviewer).

PR: https://git.openjdk.java.net/jdk/pull/1558