commit 699f6de80a312c201abf729f7c5e190425b0eac2
author John Stiles <johnstiles@google.com> Fri Oct 27 11:03:59 2023 -0400
committer SkCQ <skcq-be@skia-corp.google.com.iam.gserviceaccount.com> Fri Oct 27 15:21:37 2023 +0000
tree 5b33b946d69090c7875366c5791cb53fff59018b
parent 374ddaa86b8bb10b777ae75cd49b01069fe445d9

Restructure skcms ops to be more like Raster Pipeline.

This generates extremely similar code, but not entirely identical;
on M1, the instructions appear in a different order, and a handful
of stages use one fewer instruction. I wasn't able to discern a
performance impact via either skcms bench or nanobench.

Change-Id: I96198b24990e55b06f1e2f9e43f9b53c54d1d9ba
Bug: b/305974160
Reviewed-on: https://skia-review.googlesource.com/c/skcms/+/771419
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>