- 1d8e198 Revert "Optimize SkBlend by using NEON intrinsics" by Mike Klein · 8 years ago
- c55bc9a Fix out of bounds read in RP::load_tables_u16_be() by Matt Sarett · 8 years ago
- 7adde14 Optimize SkBlend by using NEON intrinsics by Joel Liang · 8 years ago
- e71b167 Attempt 3: SkRasterPipelineBlitter: support A8 by Mike Klein · 8 years ago
- 379938e Use RasterPipeline to support full precision on 16-bit RGBA pngs by Matt Sarett · 8 years ago
- 25b6083 Revert "Use RasterPipeline to support full precision on 16-bit RGBA pngs" by Matt Sarett · 8 years ago
- bb2339d Use RasterPipeline to support full precision on 16-bit RGBA pngs by Matt Sarett · 8 years ago
- d32303e disable runtime detected AVX2 raster pipelines by Mike Klein · 8 years ago
- c0a7a08 remove xbyak experiment by Mike Klein · 8 years ago
- 1523dbd SkXbyak: loop inside, only body by Mike Klein · 8 years ago
- 86d55b3 Revert "Retry "SkRasterPipelineBlitter: support A8"..." by Mike Klein · 8 years ago
- d052042 Fix Sk8f::Store4 (for HSW) by Matt Sarett · 8 years ago
- 83f532e Add a real SkXbyak bench, implement enough to run it. by Mike Klein · 8 years ago
- 9953737 Avoid SkFixed overflow in decal bitmap procs by Florin Malita · 8 years ago
- 7f71d88 SkXbyak basics by Mike Klein · 8 years ago
- f55ea6a Retry "SkRasterPipelineBlitter: support A8"... by Mike Klein · 8 years ago
- 8c8cb5b simplify by removing _d stages by Mike Klein · 8 years ago
- a2d25ec Use stack instead of malloc() for most calls to SkRasterPipeline::run(). by Mike Klein · 8 years ago
- 4a224f6 Revert "SkRasterPipelineBlitter: support A8" by Mike Klein · 8 years ago
- f44373c SkRasterPipelineBlitter: support A8 by Mike Klein · 8 years ago
- 464e6a1 Retry trim another instruction off SkRasterPipeline overhead by Mike Klein · 8 years ago
- e2b66a7 auto-generate _d stages by Mike Klein · 8 years ago
- 9ea894b Revert "trim another instruction off SkRasterPipeline overhead" by Mike Klein · 8 years ago
- e61c407 trim another instruction off SkRasterPipeline overhead by Mike Klein · 8 years ago
- 40a82bd remove MIPS DSP optimizations by Mike Klein · 8 years ago
- 2410717 Fix swapped interpretation of c and e in SkColorSpace_ICC by Matt Sarett · 8 years ago
- d1c550e Consolidate TILEX_LOW_BITS/TILEY_LOW_BITS -> EXTRACT_LOW_BITS by Florin Malita · 8 years ago
- 9752454 Added optimized sRGB/2.2 gamma stages into A2B color xform by raftias · 8 years ago
- 0d413f0 Revert "Revert "SkNx basically always is fast now."" by Mike Klein · 8 years ago
- d37d5d9 Revert "Revert "clamp to premul when reading premul sRGB"" by Mike Klein · 8 years ago
- 2e018f5 Revert "clamp to premul when reading premul sRGB" by Mike Klein · 8 years ago
- 8ba64d1 Revert "SkNx basically always is fast now." by Mike Klein · 8 years ago
- 04e10da clamp to premul when reading premul sRGB by Mike Klein · 8 years ago
- 21f7838 SkNx basically always is fast now. by Mike Klein · 8 years ago
- 4958006 funnel f16 through standard load/store/gather by Mike Klein · 8 years ago
- 0e18de9 Convert image width and height (used by tiling) to float once. by Mike Klein · 8 years ago
- b0b17d1 bicubic, attempt gazillion by Mike Klein · 8 years ago
- 2979b1a Added clamps in SkRasterPipeline's gamma stages. by raftias · 8 years ago
- 64d8be0 Make sure all the convolve functions are defined. by Mike Klein · 8 years ago
- 4adac2e Add AVX2 version of ConvolveVertically by xiangze.zhang · 8 years ago
- 886cf53 Refactor bilerp a little. by Mike Klein · 8 years ago
- cc63173 remove upper limit on number of pipeline stages by Mike Klein · 8 years ago
- e2e2ae2 Manual byte -> float conversion. by Mike Klein · 8 years ago
- 3e05671 Tricky float -> byte conversion in store_8888. by Mike Klein · 8 years ago
- 9c77ea1 SkNf_round, use it in store_565 and store_tables. by Mike Klein · 8 years ago
- 87185f7 use fma in store_8888 by Mike Klein · 8 years ago
- abf8ba3 SkColorSpaceXform bug fixes attempt 2 by Matt Sarett · 8 years ago
- f6878ba Reland "Add RasterPipeline implementation for SkColorSpaceXform" by Matt Sarett · 8 years ago
- 0c32496 Avoid creating std::function in run_pipeline(). by Mike Klein · 8 years ago
- 5476128 Added CMYK support for ICC profiles. by raftias · 8 years ago
- b6d4e13 Revert "Add RasterPipeline implementation for SkColorSpaceXform" by Brian Osman · 8 years ago
- dd19ac7 Add RasterPipeline implementation for SkColorSpaceXform by Matt Sarett · 8 years ago
- 62458a6 Revert "Added CMYK support for ICC profiles." by Mike Klein · 8 years ago
- 51c3fcd Added CMYK support for ICC profiles. by raftias · 8 years ago
- 14c8f82 move all memset() logic into blitter by Mike Klein · 8 years ago
- 21aa3d0 fold more in SkRasterPipelineBlitter by Mike Klein · 8 years ago
- babd93e constant means constant by Mike Klein · 8 years ago
- c789b61 Bring back SkRasterPipeline::run() for one-off uses. by Mike Klein · 8 years ago
- 9452898 restore sRGB memset optimization by Mike Klein · 8 years ago
- 7a14734d support a8 by Mike Klein · 8 years ago
- a0c4c34 teach MSAN about maskload by Mike Klein · 8 years ago
- f7657e9 gather_i8 by Mike Klein · 8 years ago
- 56b5079 Make gather() look and work more like load(). by Mike Klein · 8 years ago
- bddd234 document x_tail encoding by Mike Klein · 8 years ago
- c51c18f use fma() in accumulate by Mike Klein · 8 years ago
- 2c8104e SkNx_abi is unused. by Mike Klein · 8 years ago
- 729b582 Consistent naming. by Mike Klein · 8 years ago
- 2cbc33d simplify by Mike Klein · 8 years ago
- e03339a Convert blitter over to new style from_srgb, to_srgb. by Mike Klein · 8 years ago
- b04c352 Split srgb out of accum stages. by Mike Klein · 8 years ago
- 5a13011 Fix unpremul stage. by Mike Klein · 8 years ago
- e9f78b4 Guard against buggy ucrt\math.h. by Mike Klein · 8 years ago
- d5de013 Some simple pipeline refactoring. by Mike Klein · 8 years ago
- 6b77f1c 4444 and gray 8 by Mike Klein · 8 years ago
- cb5338c Support sRGB 565. by Mike Klein · 8 years ago
- cb2c12b accum_565 and accum_f16 by Mike Klein · 8 years ago
- 46e66a2 bilerp by Mike Klein · 8 years ago
- f7f883b Rearrange NN sampling to more naturally support bilerp. by Mike Klein · 8 years ago
- 43c847b Simplify shader paint alpha modulation. by Mike Klein · 8 years ago
- 2e35e8a mirror tiling by Mike Klein · 8 years ago
- d2265e5 Port convolve functions to SkOpts by xiangze.zhang · 8 years ago
- c01e7df perspective matrix by Mike Klein · 8 years ago
- b273fc4 repeat tiling by Mike Klein · 8 years ago
- 4e9bb74 Initialize all values we load. by Mike Klein · 8 years ago
- b6ab4ae Be careful about types in SkNx_neon.h. by Mike Klein · 8 years ago
- 06a65e2 Support SkImageShader in SkRasterPipeline blitter by Mike Klein · 8 years ago
- db4d406 Hook into parametric and table raster pipeline stages by Matt Sarett · 8 years ago
- a9312fd Add trace and registers stages. by Mike Klein · 8 years ago
- cfcf624 add {parametric,table}_{r,g,b} stages. by Mike Klein · 8 years ago
- fb191da More shader preliminaries / refactoring by Mike Klein · 8 years ago
- 0f91ea4 Fix up arguments passed to pipeline. by Mike Klein · 8 years ago
- af49b19 Start each pipeline with (x,y) in (dr,dg) registers for the shader. by Mike Klein · 8 years ago
- 2563601 Initial implementation of a SkColorSpace_A2B xform by raftias · 8 years ago
- 3f32135 Make SSE2/Neon convolution functions not to read extra bytes by xiangze.zhang · 8 years ago
- ad48a70 Give memset lambdas explicit names. by Mike Klein · 8 years ago
- c509341 swap_src_dst -> move_src_dst by Mike Klein · 8 years ago
- 125b2aa Detect pipelines which can compile to memsets. by mtklein · 8 years ago
- a4a4488 skrpb: evaluate color filters for constant shaders once. by mtklein · 8 years ago
- f0348c2 Implement SkNx_fma() for Sk4f on ARMv8. by Mike Klein · 8 years ago
- 6b059bd Remove sk_linear_to_srgb_noclamp(). by Mike Klein · 8 years ago