On Fri, 19 Jun 2026 09:25:12 GMT, Ferenc Rakoczi <[email protected]> wrote:
>>> I think the only remaining bit is that UseIntPolyIntrinsics should be >>> somewhere in AOTCODECACHE_CONFIGS_AARCH64_DO or even in >>> AOTCODECACHE_CONFIGS_GENERIC_DO. >> >> Aaargh, yes! How did I miss that? >> >> @ferakocz There is currently an entry for this flag in macro >> AOTCODECACHE_CONFIGS_X86_DO defined in file aotCodeCache.hpp >> >> #define AOTCODECACHE_CONFIGS_X86_DO(do_var, do_fun) \ >> . . . \ >> do_var(bool, UseIntPolyIntrinsics) \ >> . . . \ >> >> >> Since this option is now employed by both AArch64 and X86 You will need to >> delete the line from that macro and add it to the generic macro >> `AOTCODECACHE_CONFIGS_GENERIC_DO` -- probably best directly after the entry >> for `UsePoly1305Intrinsics`: >> >> #define AOTCODECACHE_CONFIGS_GENERIC_DO(do_var, do_fun) \ >> . . . \ >> do_var(bool, UsePoly1305Intrinsics) \ >> do_var(bool, UseIntPolyIntrinsics) \ >> . . . \ >> >> When @shipilev says GHA run he means Github Actions i.e. the automatically >> run checks and tests that are currently showing Windows failures. If you >> merge the dev head into your PR and push then it should fix the issue with >> Windows. > > Thanks @adinn and @theRealAph! Could one of you also sponsor it? Hi @ferakocz, did you have a chance to run test/micro/org/openjdk/bench/javax/crypto/full/PolynomialP256Bench.java on this patch? >.../build/patched/images/jdk/bin/java -jar >.../build/patched/images/test/micro/benchmarks.jar >org.openjdk.bench.javax.crypto.full.PolynomialP256Bench.benchAssign -p >isMontBench=true I got these numbers on my local laptop macOS on m4: Patched: >PolynomialP256Bench.benchAssign true thrpt 8 10230.113 ± >146.263 ops/s Baseline: >PolynomialP256Bench.benchAssign true thrpt 8 23548.039 ± >1596.303 ops/s ------------- PR Comment: https://git.openjdk.org/jdk/pull/30941#issuecomment-4753707997
