RFR: 8310268: RISC-V: misaligned memory access in String.Compare intrinsic [v2]

[Vladimir Kempik]

On Thu, 29 Jun 2023 05:50:06 GMT, Vladimir Kempik <vkempik at openjdk.org> wrote:

>> Please review this fix. it eliminates misaligned loads in String.compare on risc-v
>> 
>> for small compares ( <= 72 bytes), the instrinsic in c2_MacroAssembler.cpp is used,
>> it reads ( in case of UU/LL) 8 bytes per loop, and at then end, it reads tail - misaligned load of last 8 bytes from the string.
>> 
>> so if string length is not 8x bytes long then last load is misaligned, also it performs read/compare of some data which already was processed.
>> 
>> I have changed that to compare only last length%8 bytes using SHORT_STRING part of intrinsic for UL/LU. But for UU/LL I have made an optimised version.
>> 
>> Thanks to optimisations for conditional branching at line [947](https://github.com/openjdk/jdk/pull/14534/files#diff-35eb1d2f1e2f0514dd46bd7fbad49ff2c87703d5a3041a6433956df00a3fe6e6R947) I’ve got no perf drop on thead ( with +AvoidUnalignedAccesses) which supports misaligned access.
>> 
>> Improvements to inflate_XX() methods gives 3-5% improvements for UL/LU cases on thead, almost no perf change on hifive.
>> 
>> for large strings, the instrinsics from stubGenerator.cpp is used
>> for UU/LL - generate_compare_long_string_same_encoding, I have just replaced misaligned ld with load_long_misaligned. Since this tail reading is not on hot path, this give some small penalty for thead when -XX:+AvoidUnalignedAccesses.
>> 
>> large LU/UL comparision is done in compare_long_string_different_encoding in sutbGenerator.cpp:
>> These changes are partially based on feilongjiang's patch, but I have changed tail reading to prevent reading past the end of string. I have observed no perf difference between feilongjiang's and my version.
>> 
>> This also enables regression test for string.Compare which previously was aarch64-only
>> 
>> Testing: tier1 and tier2 clean on hifive.
>> 
>> JMH testing, hifive:
>> before:
>> 
>> Benchmark                                   (delta)  (size)  Mode  Cnt     Score    Error  Units
>> StringCompareToDifferentLength.compareToLL        2      24  avgt    9     6.474 ±  1.475  ms/op
>> StringCompareToDifferentLength.compareToLL        2      36  avgt    9   125.823 ±  1.947  ms/op
>> StringCompareToDifferentLength.compareToLL        2      72  avgt    9    10.512 ±  0.236  ms/op
>> StringCompareToDifferentLength.compareToLL        2     128  avgt    9    13.032 ±  0.821  ms/op
>> StringCompareToDifferentLength.compareToLL        2     256  avgt    9    18.983 ±  0.318  ms/op
>> StringCompareToDifferentLength.compareToLL        2     512  avgt    9    29.925 ± ...
>
> Vladimir Kempik has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Simplify tail for shrot string compare

Ok, updated the PR, new results for hifive:

Benchmark                                   (delta)  (size)  Mode  Cnt   Score   Error  Units
StringCompareToDifferentLength.compareToLL        2       7  avgt    9   7.472 ± 0.351  ms/op
StringCompareToDifferentLength.compareToLL        2       8  avgt    9   8.228 ± 0.801  ms/op
StringCompareToDifferentLength.compareToLL        2      15  avgt    9   8.417 ± 0.735  ms/op
StringCompareToDifferentLength.compareToLL        2      24  avgt    9   9.052 ± 0.684  ms/op
StringCompareToDifferentLength.compareToLL        2      36  avgt    9   9.587 ± 0.688  ms/op
StringCompareToDifferentLength.compareToLU        2       7  avgt    9   9.315 ± 0.552  ms/op
StringCompareToDifferentLength.compareToLU        2       8  avgt    9   9.360 ± 0.467  ms/op
StringCompareToDifferentLength.compareToLU        2      15  avgt    9  11.735 ± 0.279  ms/op
StringCompareToDifferentLength.compareToLU        2      24  avgt    9  14.576 ± 0.617  ms/op
StringCompareToDifferentLength.compareToLU        2      36  avgt    9  19.208 ± 0.901  ms/op
StringCompareToDifferentLength.compareToUL        2       7  avgt    9  12.049 ± 0.559  ms/op
StringCompareToDifferentLength.compareToUL        2       8  avgt    9  11.422 ± 0.410  ms/op
StringCompareToDifferentLength.compareToUL        2      15  avgt    9  13.699 ± 0.215  ms/op
StringCompareToDifferentLength.compareToUL        2      24  avgt    9  16.343 ± 0.652  ms/op
StringCompareToDifferentLength.compareToUL        2      36  avgt    9  19.905 ± 0.494  ms/op
StringCompareToDifferentLength.compareToUU        2       7  avgt    9   7.008 ± 0.869  ms/op
StringCompareToDifferentLength.compareToUU        2       8  avgt    9   6.996 ± 0.835  ms/op
StringCompareToDifferentLength.compareToUU        2      15  avgt    9   8.033 ± 0.800  ms/op
StringCompareToDifferentLength.compareToUU        2      24  avgt    9  10.105 ± 0.591  ms/op
StringCompareToDifferentLength.compareToUU        2      36  avgt    9  11.739 ± 0.758  ms/op

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

PR Comment: https://git.openjdk.org/jdk/pull/14534#issuecomment-1612474056