[Mesa-dev] [PATCH 11/45] i965/vec4: Handle 16-bit types at type_size_xvec4
Eduardo Lima Mitev
elima at igalia.com
Mon Aug 21 07:27:21 UTC 2017
On 08/17/2017 09:23 PM, Jason Ekstrand wrote:
>
>
> On Thu, Jul 13, 2017 at 7:35 AM, Alejandro PiƱeiro <apinheiro at igalia.com
> <mailto:apinheiro at igalia.com>> wrote:
>
> These types have similar vec4 sizes as their 32-bit counterparts.
> ---
> src/intel/compiler/brw_vec4_visitor.cpp | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/src/intel/compiler/brw_vec4_visitor.cpp
> b/src/intel/compiler/brw_vec4_visitor.cpp
> index 22ee4dd..f5ae0bc 100644
> --- a/src/intel/compiler/brw_vec4_visitor.cpp
> +++ b/src/intel/compiler/brw_vec4_visitor.cpp
> @@ -583,8 +583,11 @@ type_size_xvec4(const struct glsl_type *type,
> bool as_vec4)
> case GLSL_TYPE_UINT:
> case GLSL_TYPE_INT:
> case GLSL_TYPE_FLOAT:
> + case GLSL_TYPE_HALF_FLOAT:
> case GLSL_TYPE_BOOL:
> case GLSL_TYPE_DOUBLE:
> + case GLSL_TYPE_UINT16:
> + case GLSL_TYPE_INT16:
>
>
> Maybe make these hit the unreachable() because we'll probably never
> support fp16 in vec4.
>
Yes, but currently the FS backend calls type_size_xvec4() from
fs_visitor::nir_setup_outputs(). So we still need to provide valid vec4
sizes for 16-bit types.
We could move the function somewhere else or duplicate the logic in FS,
but it is simpler to keep this patch I think. WDYT?
> --Jason
>
>
> case GLSL_TYPE_UINT64:
> case GLSL_TYPE_INT64:
> if (type->is_matrix()) {
> --
> 2.9.3
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org <mailto:mesa-dev at lists.freedesktop.org>
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
> <https://lists.freedesktop.org/mailman/listinfo/mesa-dev>
>
>
>
>
> _______________________________________________
> 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