I'm trying to perform a masked blend (on __vector
types) on a PowerPC (POWER 8).
When looking to the intrinsics (list available here) I can see a vector select, but nothing for the merge.
On x86 processors I know the intrinsic _mm256_blendv_ps
, and I'm looking for something similar.
Thanks!
Actually vec_sel
is doing exactly what I want.
result = vec_sel(result, src, mask);
I just misread the instruction set.
Sorry for the inconvenience.