Alyssa Rosenzweig 0e0825013d agx: Do more work in agx_preprocess_nir ... agx_preprocess_nir runs once per shader, whereas agx_optimize_nir runs once per variant. That means we want to do as much work as possible in agx_preprocess_nir to make shader variants as cheap as possible to compiler. So, move our standard suite of lowering and optimizing to the preprocess loop, leaving just a single (easy) trip through the optimizer for simple variant processing. Plus, we can remove variables when preprocessing, since we no longer use variables anywhere. We remove them to reduce the RAM and disk cache footprint of shader variants. Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21104>		2023-02-20 11:34:58 +00:00
..
compiler	agx: Do more work in agx_preprocess_nir	2023-02-20 11:34:58 +00:00
drm-shim	asahi: Clang-format the subtree	2022-12-27 22:46:29 +00:00
layout	ail: Test 63x63 cube map	2023-02-04 07:04:49 +00:00
lib	asahi: Vectorize background colour load	2023-02-16 06:36:49 +00:00
.clang-format	asahi: Add .clang-format file	2022-12-27 22:46:29 +00:00
meson.build	asahi: Add drm-shim implementation	2022-11-07 16:10:11 +00:00