Kristof Roomp

improve JPEG encode_block perf by about 25% with faster integer encoding

6

I license past and future contributions under the dual MIT/Apache-2.0 license, allowing licensees to choose either at their option. Also add benches for tracking performance of JPEG block encoding. pre:...

Making new const simplified a bunch of code and allows users of this library to define their own constants without having to do anything hacky. It does mean that it...

Add implementations for all the intrinsics @Melirius might be interested

Optionally use wide crate for DCT, which adds support for ARM64, WASM and SSE2 on stable

7

The result is a 50% speedup for non-AVX SSE2 (default for rust x86_64). If compiling for AVX2, the code is about 20% slower than the hand-rolled AVX2 code. The wide...

add alignment on JPEG [i16; 64] blocks

5

Ensure that we access coefficient blocks 32 byte aligned so that we can autovectorize and optimize some codepaths via SIMD

I tried using multiversion crate for an avx2 version of a method that was using wide. It seems to work and get selected properly, was there some reason it shouldn't?

Potential feature: safe align_to

4

What do you think of something like this? I realize that you can also do this with bytemuck, but this would let you write more generic code when operating on...