On Fri, 5 Apr 2024 21:53:50 GMT, Scott Gibbons <sgibb...@openjdk.org> wrote:
>> This code makes an intrinsic stub for `Unsafe::setMemory` for x86_64. See >> [this PR](https://github.com/openjdk/jdk/pull/16760) for discussion around >> this change. >> >> Overall, making this an intrinsic improves overall performance of >> `Unsafe::setMemory` by up to 4x for all buffer sizes. >> >> Tested with tier-1 (and full CI). I've added a table of the before and >> after numbers for the JMH I ran (`MemorySegmentZeroUnsafe`). >> >> [setMemoryBM.txt](https://github.com/openjdk/jdk/files/14808974/setMemoryBM.txt) > > Scott Gibbons has updated the pull request incrementally with one additional > commit since the last revision: > > Fixed generate_fill when count > 0x80000000 Thanks for all your thoughtful feedback. I would very much like to take "the right approach"(tm) but I don't have the skill to write IR, especially given that this is an Unsafe block, which is restricted by atomicity and alignment. I would not know how to prevent the C2 optimizer from vectorizing, or indeed replacing my code with a call to memset(). I'm not sure it would go this far, but in order to remain compliant with the spec I have to prevent it in the future. This was modeled after the existing implementation of copyMemory, gives good performance (3-5x), and can serve as a template for other platform developers to follow. They have the expertise for their specific platform(s) which I do not have. Again, thank you. ------------- PR Comment: https://git.openjdk.org/jdk/pull/18555#issuecomment-2040688225
