projects / platform / upstream / mesa.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 7068154) | patch
agx: Implement extract_[ui]16
authorAlyssa Rosenzweig <alyssa@rosenzweig.io>
Fri, 3 Mar 2023 05:29:49 +0000 (00:29 -0500)
committerMarge Bot <emma+marge@anholt.net>
Sat, 11 Mar 2023 14:15:50 +0000 (14:15 +0000)
commit8a53050d7d3fd800c211e20fe9d9f83249ca098c
tree10ad76ece62d0e5236e31151cefe44c39275ccf7tree | snapshot
parent706815488e19114437f3bf3c2c7672dc4fd7e6adcommit | diff
agx: Implement extract_[ui]16

Instead of lowering to bitwise ops. Yet another way of subdividing in NIR.
Probably insignificant but makes it easy to check that the pass ordering from the
previous pass is right. It does let us get much better codegen for
unpacksnorm2x16, whatever that's worth.

No shader-db changes.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21674>
src/asahi/compiler/agx_compile.h diff | blob | history
src/asahi/compiler/agx_nir_algebraic.py diff | blob | history
Domain: Graphics System / GL;