RFR: 8322770: Implement C2 VectorizedHashCode on AArch64 [v10]
Andrew Haley
aph at openjdk.org
Fri Sep 20 09:53:45 UTC 2024
On Thu, 19 Sep 2024 15:05:04 GMT, Mikhail Ablakatov <duke at openjdk.org> wrote:
>> Hello,
>>
>> Please review the following PR for [JDK-8322770 Implement C2 VectorizedHashCode on AArch64](https://bugs.openjdk.org/browse/JDK-8322770). It follows previous work done in https://github.com/openjdk/jdk/pull/16629 and https://github.com/openjdk/jdk/pull/10847 for RISC-V and x86 respectively.
>>
>> The code to calculate a hash code consists of two parts: a vectorized loop of Neon instruction that process 4 or 8 elements per iteration depending on the data type and a fully unrolled scalar "loop" that processes up to 7 tail elements.
>>
>> At the time of writing this I don't see potential benefits from providing SVE/SVE2 implementation, but it could be added as a follow-up or independently later if required.
>>
>> # Performance
>>
>> ## Neoverse N1
>>
>>
>> --------------------------------------------------------------------------------------------
>> Version Baseline This patch
>> --------------------------------------------------------------------------------------------
>> Benchmark (size) Mode Cnt Score Error Score Error Units
>> --------------------------------------------------------------------------------------------
>> ArraysHashCode.bytes 1 avgt 15 1.249 ? 0.060 1.247 ? 0.062 ns/op
>> ArraysHashCode.bytes 10 avgt 15 8.754 ? 0.028 4.387 ? 0.015 ns/op
>> ArraysHashCode.bytes 100 avgt 15 98.596 ? 0.051 26.655 ? 0.097 ns/op
>> ArraysHashCode.bytes 10000 avgt 15 10150.578 ? 1.352 2649.962 ? 216.744 ns/op
>> ArraysHashCode.chars 1 avgt 15 1.286 ? 0.062 1.246 ? 0.054 ns/op
>> ArraysHashCode.chars 10 avgt 15 8.731 ? 0.002 5.344 ? 0.003 ns/op
>> ArraysHashCode.chars 100 avgt 15 98.632 ? 0.048 23.023 ? 0.142 ns/op
>> ArraysHashCode.chars 10000 avgt 15 10150.658 ? 3.374 2410.504 ? 8.872 ns/op
>> ArraysHashCode.ints 1 avgt 15 1.189 ? 0.005 1.187 ? 0.001 ns/op
>> ArraysHashCode.ints 10 avgt 15 8.730 ? 0.002 5.676 ? 0.001 ns/op
>> ArraysHashCode.ints 100 avgt 15 98.559 ? 0.016 24.378 ? 0.006 ns/op
>> ArraysHashCode.ints 10000 avgt 15 10148.752 ? 1.336 2419.015 ? 0.492 ns/op
>> ArraysHashCode.multibytes 1 avgt 15 1.037 ? 0.001 1.037 ? 0.001 ...
>
> Mikhail Ablakatov has updated the pull request incrementally with four additional commits since the last revision:
>
> - cleanup: use switch-case instead of if-else statements and ternary operators
> - Don't try align basic blocks as it brings no measurable performance benefits
> - fixup: rename the newly added Vector-Scalar mulv to mulvs
> - fixup: fix Windows build by not using RELATIVE as an identifier
src/hotspot/cpu/aarch64/c2_MacroAssembler_aarch64.cpp line 140:
> 138: }
> 139:
> 140: void C2_MacroAssembler::arrays_hashcode_elload(Register dst, Address src, BasicType eltype) {
This method has nothing to do with either arrays nor hashcode. Looking at class `Assembler`, it would make sense to have a general-purpose load that takes a `BasicType`.
We already have this `Assembler` function:
` void ld_st2(Register Rt, const Address &adr, int size, int op, int V = 0)`
where two bits of `op` gives us store and three versions of sign extension.
Please use `ld_st2` to define a general-purpose `MacroAssembler::load` function with this set of arguments.
It was perhaps a historic mistake of ours not to use an assembler function with the size and the signedness of an operand as an argument to ldr/str.
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/18487#discussion_r1768320227
More information about the hotspot-dev
mailing list