[Mesa-dev] [PATCH 03/59] freedreno/ir3: lower lrp when operating with double operands

Rob Clark robdclark at gmail.com
Wed May 4 11:52:00 UTC 2016


On Fri, Apr 29, 2016 at 7:29 AM, Samuel Iglesias Gonsálvez
<siglesias at igalia.com> wrote:
> Lower lrp when operating with double operands because float version of
> lrp is also lowered.
>
> Signed-off-by: Samuel Iglesias Gonsálvez <siglesias at igalia.com>
> CC: Rob Clark <robdclark at gmail.com>
> ---
>  src/gallium/drivers/freedreno/ir3/ir3_nir.c | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/src/gallium/drivers/freedreno/ir3/ir3_nir.c b/src/gallium/drivers/freedreno/ir3/ir3_nir.c
> index 364e92b..0635dfb 100644
> --- a/src/gallium/drivers/freedreno/ir3/ir3_nir.c
> +++ b/src/gallium/drivers/freedreno/ir3/ir3_nir.c
> @@ -43,6 +43,7 @@ ir3_tgsi_to_nir(const struct tgsi_token *tokens)
>                         .lower_fsat = true,
>                         .lower_scmp = true,
>                         .lower_flrp32 = true,
> +                       .lower_flrp64 = true,
>                         .lower_ffract = true,
>                         .native_integers = true,
>                         .vertex_id_zero_based = true,

I guess there are probably other things missing for fp64 support, but
I think we probably want to figure that out at some point.  According
to opencl level supported by a4xx+ it should support doubles, but I'm
not sure yet if that is just via lowering.  Anyways, I guess this
makes sense for now, so:

Reviewed-by: Rob Clark <robdclark at gmail.com>



> --
> 2.5.0
>
> _______________________________________________
> 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