[Mesa-dev] [PATCH 2/3] swr/rast: don't use _mm256_fmsub_ps in AVX code

Cherniak, Bruce bruce.cherniak at intel.com
Thu Jul 6 19:38:04 UTC 2017


Reviewed-by: Bruce Cherniak <bruce.cherniak at intel.com>

> On Jul 6, 2017, at 11:26 AM, Tim Rowley <timothy.o.rowley at intel.com> wrote:
> 
> ---
> src/gallium/drivers/swr/rasterizer/common/simdlib_256_avx.inl | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/src/gallium/drivers/swr/rasterizer/common/simdlib_256_avx.inl b/src/gallium/drivers/swr/rasterizer/common/simdlib_256_avx.inl
> index aec79e3..16eb521 100644
> --- a/src/gallium/drivers/swr/rasterizer/common/simdlib_256_avx.inl
> +++ b/src/gallium/drivers/swr/rasterizer/common/simdlib_256_avx.inl
> @@ -182,7 +182,11 @@ static SIMDINLINE Float SIMDCALL fmadd_ps(Float a, Float b, Float c) // return (
>     return add_ps(mul_ps(a, b), c);
> }
> 
> -SIMD_WRAPPER_3(fmsub_ps);   // return (a * b) - c
> +static SIMDINLINE Float SIMDCALL fmsub_ps(Float a, Float b, Float c) // return (a * b) - c
> +{
> +    return sub_ps(mul_ps(a, b), c);
> +}
> +
> SIMD_WRAPPER_2(max_ps);     // return (a > b) ? a : b
> SIMD_WRAPPER_2(min_ps);     // return (a < b) ? a : b
> SIMD_WRAPPER_2(mul_ps);     // return a * b
> -- 
> 2.7.4
> 
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev



More information about the mesa-dev mailing list