https://bugs.winehq.org/show_bug.cgi?id=49663
--- Comment #11 from Zebediah Figura z.figura12@gmail.com --- (In reply to Fabian Maurer from comment #10)
The glibc implementation offers about up to 50% more speed, but it's all written in assembly to take advantage of AVX/SSE, combined with architecture checks to choose the best path at runtime. I didn't want to replicate all that into Wine, when musl is fast enough. What's your thought on that?
While the reporter's numbers are not necessarily statistically significant on their own, generally 20 fps is an improvement worth having. Personally, I think when it comes to core CRT functions like memmove, it really is worth having as optimized an implementation as possible. I don't get the impression that platform-specific assembly would be much of a problem.