RFR: 8332587: RISC-V: secondary_super_cache does not scale well [v9]

[Gui Cao]

On Thu, 20 Jun 2024 02:20:18 GMT, Fei Yang <fyang at openjdk.org> wrote:

>> Gui Cao has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Add population_count for scalar version
>
> src/hotspot/cpu/riscv/macroAssembler_riscv.cpp line 3635:
> 
>> 3633:       addi(dst, dst, 1);
>> 3634:       mv(tmp2, tmp1);
>> 3635:       addi(tmp2, tmp2, -1);
> 
> Suggestion: `addi(tmp2, tmp1, -1);`. This help save the preceding `mv` instruction.

Thanks for the review. I've retested the JMH.

JMH tested on SOPHON SG2042 (has not Zbb)
with patch:

Benchmark                             Mode  Cnt    Score   Error  Units
SecondarySupersLookup.testNegative00  avgt   15   20.649 ± 0.147  ns/op
SecondarySupersLookup.testNegative01  avgt   15   20.649 ± 0.117  ns/op
SecondarySupersLookup.testNegative02  avgt   15   20.637 ± 0.116  ns/op
SecondarySupersLookup.testNegative03  avgt   15   20.638 ± 0.113  ns/op
SecondarySupersLookup.testNegative04  avgt   15   20.638 ± 0.127  ns/op
SecondarySupersLookup.testNegative05  avgt   15   20.639 ± 0.115  ns/op
SecondarySupersLookup.testNegative06  avgt   15   20.638 ± 0.119  ns/op
SecondarySupersLookup.testNegative07  avgt   15   20.850 ± 0.457  ns/op
SecondarySupersLookup.testNegative08  avgt   15   20.842 ± 0.459  ns/op
SecondarySupersLookup.testNegative09  avgt   15   20.650 ± 0.124  ns/op
SecondarySupersLookup.testNegative10  avgt   15   20.642 ± 0.127  ns/op
SecondarySupersLookup.testNegative16  avgt   15   20.657 ± 0.157  ns/op
SecondarySupersLookup.testNegative20  avgt   15   20.669 ± 0.152  ns/op
SecondarySupersLookup.testNegative30  avgt   15   20.668 ± 0.166  ns/op
SecondarySupersLookup.testNegative32  avgt   15   20.669 ± 0.168  ns/op
SecondarySupersLookup.testNegative40  avgt   15   20.668 ± 0.174  ns/op
SecondarySupersLookup.testNegative50  avgt   15   20.682 ± 0.194  ns/op
SecondarySupersLookup.testNegative55  avgt   15  113.369 ± 3.792  ns/op
SecondarySupersLookup.testNegative56  avgt   15  113.888 ± 3.769  ns/op
SecondarySupersLookup.testNegative57  avgt   15  115.320 ± 4.271  ns/op
SecondarySupersLookup.testNegative58  avgt   15  115.648 ± 2.985  ns/op
SecondarySupersLookup.testNegative59  avgt   15  117.730 ± 3.370  ns/op
SecondarySupersLookup.testNegative60  avgt   15  142.533 ± 3.636  ns/op
SecondarySupersLookup.testNegative61  avgt   15  144.901 ± 5.267  ns/op
SecondarySupersLookup.testNegative62  avgt   15  145.926 ± 3.799  ns/op
SecondarySupersLookup.testNegative63  avgt   15  207.704 ± 5.370  ns/op
SecondarySupersLookup.testNegative64  avgt   15  210.631 ± 3.832  ns/op
SecondarySupersLookup.testPositive01  avgt   15   20.334 ± 0.455  ns/op
SecondarySupersLookup.testPositive02  avgt   15   20.126 ± 0.101  ns/op
SecondarySupersLookup.testPositive03  avgt   15   20.126 ± 0.097  ns/op
SecondarySupersLookup.testPositive04  avgt   15   20.124 ± 0.102  ns/op
SecondarySupersLookup.testPositive05  avgt   15   20.119 ± 0.100  ns/op
SecondarySupersLookup.testPositive06  avgt   15   20.126 ± 0.098  ns/op
SecondarySupersLookup.testPositive07  avgt   15   20.321 ± 0.462  ns/op
SecondarySupersLookup.testPositive08  avgt   15   20.117 ± 0.098  ns/op
SecondarySupersLookup.testPositive09  avgt   15   20.534 ± 0.555  ns/op
SecondarySupersLookup.testPositive10  avgt   15   20.120 ± 0.100  ns/op
SecondarySupersLookup.testPositive16  avgt   15   20.125 ± 0.104  ns/op
SecondarySupersLookup.testPositive20  avgt   15   20.125 ± 0.116  ns/op
SecondarySupersLookup.testPositive30  avgt   15   20.132 ± 0.110  ns/op
SecondarySupersLookup.testPositive32  avgt   15   20.328 ± 0.449  ns/op
SecondarySupersLookup.testPositive40  avgt   15   20.132 ± 0.096  ns/op
SecondarySupersLookup.testPositive50  avgt   15   20.331 ± 0.460  ns/op
SecondarySupersLookup.testPositive60  avgt   15   20.134 ± 0.104  ns/op
SecondarySupersLookup.testPositive63  avgt   15   20.128 ± 0.104  ns/op
SecondarySupersLookup.testPositive64  avgt   15   20.334 ± 0.456  ns/op
Finished running test 'micro:vm.lang.SecondarySupersLookup'


The JMH test data is a little better than before, Thanks a lot.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/19320#discussion_r1646875539