OpenHarmony/third_party_mesa3d
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

intel/compiler: force simd8 when dual src blending on gen8

Browse Source 
Patch introduces option to force simd8 and uses it as a workaround for
dual source blending issues seen with skqp (skia testsuite) on gen8.

Fixes following Piglit test on gen8 platforms:
   arb_blend_func_extended-dual-src-blending-issue-1917

Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/1917
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
c: <mesa-stable@lists.freedesktop.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
This commit is contained in:
Tapani Pälli
2019-12-02 16:54:30 +02:00
parent f6004bac1f
commit 32ebd4207a
1 changed files with 10 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
src/intel/compiler/brw_fs.cpp
View File

@@ -8274,6 +8274,16 @@ brw_compile_fs(const struct brw_compiler *compiler, void *log_data,
prog_data->reg_blocks_8 = brw_register_blocks(v8.grf_used);
}
/* Limit dispatch width to simd8 with dual source blending on gen8.
* See: https://gitlab.freedesktop.org/mesa/mesa/issues/1917
*/
if (devinfo->gen == 8 && prog_data->dual_src_blend &&
!(INTEL_DEBUG & DEBUG_NO8)) {
assert(!use_rep_send);
v8.limit_dispatch_width(8, "gen8 workaround: "
"using SIMD8 when dual src blending.\n");
}
if (v8.max_dispatch_width >= 16 &&
likely(!(INTEL_DEBUG & DEBUG_NO16) || use_rep_send)) {
/* Try a SIMD16 compile */