#Is it possible to use arch-specific SIMD intrinsics directly from Zig?
1 messages · Page 1 of 1 (latest)
you can always do inline asm
if the builtin vector types aren't suitable
I suppose you could try to @cImport an intrinsics header - I just did inline asm when I was vectorising things in the past
I'd rather not do register allocation myself 😄
does that need libc?
yes, in some cases they work nice, in some cases they don't
but I will try importing the intrinsics header