[Mesa-dev] [PATCH 2/2] llvmpipe: turn depth clears into full depth/stencil clears for d24x8 formats
Jose Fonseca
jfonseca at vmware.com
Mon Jan 18 14:55:35 PST 2016
On 18/01/16 03:33, sroland at vmware.com wrote:
> From: Roland Scheidegger <sroland at vmware.com>
>
> If we have a d24x8 format, there is no stencil. Therefore, we can always
> clear these bits too, which means this will be some kind of memset rather
> than read-modify-write.
> This is good for some 7% increase or so in gears with huge window size -
> seems to have a bigger effect if things aren't in caches. Of course, any
> real app won't spend nearly as much time comparatively in clearing
> depth buffer in the first place, so the speedup will be much lower.
> ---
> src/gallium/drivers/llvmpipe/lp_setup.c | 25 ++++++++++++++-----------
> 1 file changed, 14 insertions(+), 11 deletions(-)
>
> diff --git a/src/gallium/drivers/llvmpipe/lp_setup.c b/src/gallium/drivers/llvmpipe/lp_setup.c
> index e8c3e7c..34d3c81 100644
> --- a/src/gallium/drivers/llvmpipe/lp_setup.c
> +++ b/src/gallium/drivers/llvmpipe/lp_setup.c
> @@ -476,27 +476,30 @@ lp_setup_try_clear_zs(struct lp_setup_context *setup,
> uint64_t zsvalue = 0;
> uint32_t zmask32;
> uint8_t smask8;
> + enum pipe_format format = setup->fb.zsbuf->format;
>
> LP_DBG(DEBUG_SETUP, "%s state %d\n", __FUNCTION__, setup->state);
>
> zmask32 = (flags & PIPE_CLEAR_DEPTH) ? ~0 : 0;
> smask8 = (flags & PIPE_CLEAR_STENCIL) ? ~0 : 0;
>
> - zsvalue = util_pack64_z_stencil(setup->fb.zsbuf->format,
> - depth,
> - stencil);
> + zsvalue = util_pack64_z_stencil(format, depth, stencil);
>
> - /*
> - * XXX: should make a full mask here for things like D24X8,
> - * otherwise we'll do a read-modify-write clear later which
> - * should be unnecessary.
> - */
> - zsmask = util_pack64_mask_z_stencil(setup->fb.zsbuf->format,
> - zmask32,
> - smask8);
> + zsmask = util_pack64_mask_z_stencil(format, zmask32, smask8);
>
> zsvalue &= zsmask;
>
> + if (format == PIPE_FORMAT_Z24X8_UNORM ||
> + format == PIPE_FORMAT_X8Z24_UNORM) {
> + /*
> + * Make full mask if there's "X" bits so we can do full
> + * clear (without rmw).
> + */
> + uint32_t zsmask_full = 0;
> + zsmask_full = util_pack_mask_z_stencil(format, ~0, ~0);
> + zsmask |= ~zsmask_full;
> + }
> +
> if (setup->state == SETUP_ACTIVE) {
> struct lp_scene *scene = setup->scene;
>
>
Series is
Reviewed-by: Jose Fonseca <jfonseca at vmware.com>
More information about the mesa-dev
mailing list