Use transpose op to generate diagonal matrices.
We already had a working implementation here, but the instruction
count savings here were too good to pass up.
This CL also generalizes transposing and swizzling logic. This will
make it easier to design ops that utilize `transpose` in novel ways,
and also causes 2x2 transposes to automatically become swizzle-4s.
Change-Id: I3c396ef0b7c9c3f4f75ea12983b9dea362d90fe1
Bug: skia:13676
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/629437
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
4 files changed
