[llvm] [CodeGen][PreISelIntrinsicLowering] Add VP-based lowering for memcpy/memmove/memset (PR #165585)

[David Del Río via llvm-commits]

dadra-oc wrote:

Thanks for the review.

That's fair. To be honest, I implemented it with RISC-V target in mind, but I thought that it could be also interesting for AArch64. If you think its not worth to have it for AArch64 I can just restrict this feature for RISC-V. 

In any case, I also wanted to be conservative in this sense and this will only apply on the case the calls to those library functions are not available in TLI or if the option flag is used.




https://github.com/llvm/llvm-project/pull/165585