RFR: 8339983: [s390x] secondary_super_cache does not scale well: C1 and interpreter

[Amit Kumar]

On Wed, 27 Nov 2024 16:03:55 GMT, Lutz Schmidt <lucy at openjdk.org> wrote:

>> On 11\/27\/24 15\:58\, Amit Kumar wrote\:
>>> On Wed\, 27 Nov 2024 15\:48\:29 GMT\, Lutz Schmidt \<lucy at openjdk\.org> wrote\:
>>> 
>>>> Are there any performance figures\? Does the change help\, or at least do no harm\, on s390\?
>>> 
>>> I do but they don\'t look interesting\.
>> 
>> With the patch you should see this\:
>> 
>> \-XX\:TieredStopAtLevel\=1 \-XX\:\+UnlockDiagnosticVMOptions \-XX\:\+UseSecondarySupersCache \-XX\:\-UseSecondarySupersTable
>> 
>> Benchmark                                   Mode  Cnt    Score     Error  Units
>> SecondarySuperCacheInterContention\.test     avgt    5  176\.328 \?  69\.341  ns\/op
>> SecondarySuperCacheInterContention\.test\:t1  avgt    5  173\.024 \?  29\.227  ns\/op
>> SecondarySuperCacheInterContention\.test\:t2  avgt    5  179\.633 \? 110\.211  ns\/op
>> 
>> \-XX\:TieredStopAtLevel\=1
>> 
>> Benchmark                                   Mode  Cnt  Score   Error  Units
>> SecondarySuperCacheInterContention\.test     avgt    5  5\.156 \? 0\.855  ns\/op
>> SecondarySuperCacheInterContention\.test\:t1  avgt    5  5\.225 \? 1\.604  ns\/op
>> SecondarySuperCacheInterContention\.test\:t2  avgt    5  5\.087 \? 0\.740  ns\/op
>> 
>> assuming you have >1 core\. Try also SecondarySuperCacheIntraContention\.test
>> 
>> \-\- 
>> Andrew Haley  \(he\/him\)
>> Java Platform Lead Engineer
>> Red Hat UK Ltd\. \<https\:\/\/www\.redhat\.com>
>> https\:\/\/keybase\.io\/andrewhaley
>> EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671
>
>> I do but they don't look interesting.
> 
> OK. No harm detected.

Thank you for reviews and suggestions @RealLucy, @theRealAph.

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

PR Comment: https://git.openjdk.org/jdk/pull/22341#issuecomment-2513502558