diff options
author | mtklein <mtklein@chromium.org> | 2015-07-22 10:52:53 -0700 |
---|---|---|
committer | Commit bot <commit-bot@chromium.org> | 2015-07-22 10:52:53 -0700 |
commit | ced1585149d4ac8a68efd80e11dbc23bca6620d4 (patch) | |
tree | ca42ea0d7f2d818dbf15a0cb83a9a934a032664d /src/opts/Sk4px_none.h | |
parent | 651c920d2f8671756968366c7ed9b455fc1a096c (diff) |
565 support for SIMD xfermodes
This uses the most basic approach possible:
- to load an Sk4px from 565, convert to SkPMColors on the stack serially then load those SkPMColors.
- to store an Sk4px to 565, store to SkPMColors on the stack then convert to 565 serially.
Clearly, we can optimize these loads and stores. That's a TODO.
The code using SkPMFloat is the same idea but a little more long-term viable, as we're only operating on one pixel at a time anyway. We could probably write 565 <-> SkPMFloat methods, but I'd rather not until it's really compelling.
The speedups are varied but similar across SSE and NEON: a few uninteresting, many 50% faster, some 2x faster, and SoftLight ~4x faster.
This will cause minor GM diffs, but I don't think any layout test changes.
BUG=skia:
Committed: https://skia.googlesource.com/skia/+/942930dcaa51f66d82cdaf46ae62efebd16c8cd0
Committed: https://skia.googlesource.com/skia/+/860dcaa2ddfdadc050af4f943a84a9d499315066
Review URL: https://codereview.chromium.org/1245673002
Diffstat (limited to 'src/opts/Sk4px_none.h')
-rw-r--r-- | src/opts/Sk4px_none.h | 31 |
1 files changed, 31 insertions, 0 deletions
diff --git a/src/opts/Sk4px_none.h b/src/opts/Sk4px_none.h index ba13e58fb5..540edb821d 100644 --- a/src/opts/Sk4px_none.h +++ b/src/opts/Sk4px_none.h @@ -100,4 +100,35 @@ inline Sk4px Sk4px::zeroColors() const { 0,0,0, this->kth<15>()); } +inline Sk4px Sk4px::Load4(const SkPMColor16 src[4]) { + SkPMColor src32[4]; + for (int i = 0; i < 4; i++) { src32[i] = SkPixel16ToPixel32(src[i]); } + return Load4(src32); +} +inline Sk4px Sk4px::Load2(const SkPMColor16 src[2]) { + SkPMColor src32[2]; + for (int i = 0; i < 2; i++) { src32[i] = SkPixel16ToPixel32(src[i]); } + return Load2(src32); +} +inline Sk4px Sk4px::Load1(const SkPMColor16 src[1]) { + SkPMColor src32 = SkPixel16ToPixel32(src[0]); + return Load1(&src32); +} + +inline void Sk4px::store4(SkPMColor16 dst[4]) const { + SkPMColor dst32[4]; + this->store4(dst32); + for (int i = 0; i < 4; i++) { dst[i] = SkPixel32ToPixel16(dst32[i]); } +} +inline void Sk4px::store2(SkPMColor16 dst[2]) const { + SkPMColor dst32[2]; + this->store2(dst32); + for (int i = 0; i < 2; i++) { dst[i] = SkPixel32ToPixel16(dst32[i]); } +} +inline void Sk4px::store1(SkPMColor16 dst[1]) const { + SkPMColor dst32; + this->store1(&dst32); + dst[0] = SkPixel32ToPixel16(dst32); +} + } // namespace |