On Mon, 21 Oct 2024 18:18:12 GMT, Vladimir Ivanov <vaiva...@openjdk.org> wrote:
> This patch remove access to the shared variable to fix scalability issue in > the multithread environment. According to testing by the > specjvm2008::crypto.rsa the one thread performance reduced for less than 1% > while the score for the multithread run increased in ~2x. For the 2 socket > system with Xeon 8480+ numbers looks as: > - 1 thread: 643.15 for original version vs 642.54 for patched one; > - 224 threads: 22446.19 for original vs 46147.41 for patched. > > The RSABench microbenchmark reports no score changes for the 1 thread > (average for all testcases) and 2.4% improvement for the 224 threads. I was failed to find any real application that was affected. For now only specjvm2008 reports different scores. Seems, this optimization was designed for the one thread performance long time ago but now even for one thread it gives less than run-to-run variation. ------------- PR Comment: https://git.openjdk.org/jdk/pull/21622#issuecomment-2427942627
