RFR: 8322770: Implement C2 VectorizedHashCode on AArch64 [v5]

[Andrew Dinn]

On Wed, 21 Aug 2024 16:11:25 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 one additional commit since the last revision:
> 
>   cleanup: use a constexpr function for intpow instead of a templated class

src/hotspot/cpu/aarch64/c2_MacroAssembler_aarch64.cpp line 100:

> 98: 
> 99:   assert(is_power_of_2(unroll_factor), "can't use this value to calculate the jump target PC");
> 100:   orr(tmp2, cnt, 0x1fff ^ (unroll_factor - 1));

This is rather cryptic, especially the slightly arbitrary choice of 0x1fff -- any all 1s value greater than (loop_factor - 1) would do. I think a comment as to what is being done here might help maintainers understand what is happening here.

// At this point cnt holds (r - l) where r is the number of remaining elements,
// l is loop_count and 0 <= r < l. The constant operand to the orr 'rounds' this
// negative value into range [-u, -1] where u = unroll_factor, by clearing bits
// below bit k = log2(unroll_factor) and setting any higher bits that might be clear.
// The subtract shifted by 3 offsets past ((u - r) % u) load + madd insns i.e. it only
// executes r % u load + madds. Iteration eats up the remainder, u elements at a time.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/18487#discussion_r1727033707