projects / platform / upstream / mesa.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 2cf93f7) | patch
intel/fs: optimize uniform SSBO & shared loads
authorLionel Landwerlin <lionel.g.landwerlin@intel.com>
Mon, 6 Mar 2023 18:11:53 +0000 (20:11 +0200)
committerMarge Bot <emma+marge@anholt.net>
Wed, 5 Apr 2023 12:32:56 +0000 (12:32 +0000)
commita358b97c58699a43cc0acf9120ab4476e4a281bd
treeab7d4ec8bc07e451e6ff0dde89f24f105cae83fatree | snapshot
parent2cf93f76322331176256881d510160ec4bcf718acommit | diff
intel/fs: optimize uniform SSBO & shared loads

Using divergence analysis, figure out when SSBO & shared memory loads
are uniform and carry the data only once in register space.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21853>
src/intel/compiler/brw_fs_nir.cpp diff | blob | history
src/intel/compiler/brw_nir.c diff | blob | history
src/intel/compiler/brw_nir.h diff | blob | history
src/intel/compiler/brw_nir_blockify_uniform_loads.c [new file with mode: 0644] blob
src/intel/compiler/meson.build diff | blob | history
Domain: Graphics System / GL;