Alexander Karpinsky comments

Results 93 comments of


                                            Alexander Karpinsky

Optimization encoding in 124 times

Also back ported to upstream: https://github.com/woltapp/blurhash/pull/256

Optimization encoding in 124 times

I have updated branch to match the latest version from https://github.com/woltapp/blurhash/pull/256 @lautat Are you interested in this improvements?

Speedup C encoder up to 100x

~~I've also implemented SSE and NEON optimizations in [separate branch](https://github.com/homm/blurhash/blob/optimize-simd/C/encode.c).~~ The last optimization with unrolling loop in `multiplyBasisFunction` is actually works better since it allows any compiler effectively autovectorize the...

Speedup C encoder up to 100x

@DagAgren Are you interested in this improvements?

Speedup C encoder up to 100x

I also improved decoder performance about 14 times using the same techniques: caching cos values, linearTosRGB values and unrolling loops. This improves performance of decoding from 6 Mpx/s to 86...

Speedup C encoder up to 100x

@DagAgren How can I earn your attention?

Speedup C encoder up to 100x

> However, this code is written intentionally to be simple rather than performant Does this mean you’re rejecting any performance improvements entirely, or only the more radical ones (like 4×...

GaussianBlur does not work with I;16

> split 10-bit 3-channel images to 3 10-bit 1-channel images in advance, then use Pillow to handle the LUT Again, this wouldn't 3D LUT, this will be three 1D LUTs.

Allow using SIMD intrinsics for SSE2, SSE4, AVX2 and NEON code

@python-pillow/pillow-team What do you think?

Optimize getbbox() and getextrema() routines

@radarhere Ok, added, but it's just some random image from stocks)