You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have a use case which makes heavy use of integer vectors, which lead me to notice that there are no UVec3A and IVec3A types, and that the UVec4 and IVec4 types do not actually have SIMD implementations. Sorry if I have missed something and that there are in fact SIMD implementations for them, but if not, do you have plans to add them?
If not, I will work on a PR to add them if you'd want those implementations in the library.
The text was updated successfully, but these errors were encountered:
Hi there. You are correct, there are no integer SIMD implementations at the moment.
I am in two minds about adding support. For one, it is quite a lot of work - especially to support all of the architectures that glam currently supports (although this isn't a blocker, missing implements have a scalar path). Another thing I have found when adding support for NEON is the benefits have been less clear than they were with SSE2. I am considering changing how glam supports SIMD - mostly just making it more opt in than it currently is for Vec4/Quat and Mat4, that wouldn't negate any integer SIMD work though, it would just mean that it moves. In any case if you are interested in doing the work then I would support a PR, but as I don't have a huge amount of time to spend on this myself it would need to be fairly complete.
Hi,
Thanks for the awesome work!
I have a use case which makes heavy use of integer vectors, which lead me to notice that there are no
UVec3A
andIVec3A
types, and that theUVec4
andIVec4
types do not actually have SIMD implementations. Sorry if I have missed something and that there are in fact SIMD implementations for them, but if not, do you have plans to add them?If not, I will work on a PR to add them if you'd want those implementations in the library.
The text was updated successfully, but these errors were encountered: