Hello Richard,
On 14 фев 11:26, Richard Biener wrote:
>
> The following tries to account for the fact that when constructing
> AVX256 or AVX512 vectors from elements we can only use insertps to
> insert into the low 128bits of a vector but have to use
> vinserti128 or vinserti64x4 to build larger
][i386] Adjust vec_construct cost for AVX256/512, penaltize
elementwise load vectorization
The following tries to account for the fact that when constructing
AVX256 or AVX512 vectors from elements we can only use insertps to insert into
the low 128bits of a vector but have to use
vinse
The following tries to account for the fact that when constructing
AVX256 or AVX512 vectors from elements we can only use insertps to
insert into the low 128bits of a vector but have to use
vinserti128 or vinserti64x4 to build larger AVX256/512 vectors.
Those operations also have higher latency