Nathan Moinvaziri
Nathan Moinvaziri
> I just had a renewed look at this, and considering the earlier unaligned debate, would it perhaps be better if we removed the unaligned_16 variant of this but kept...
Updated `compare256_rle` benchmarks: ``` Run on (20 X 2117.83 MHz CPU s) CPU Caches: L1 Data 48 KiB (x10) L1 Instruction 32 KiB (x10) L2 Unified 1280 KiB (x10) L3...
I wonder if it can be optimized further for the /1 cases.
Instead of doing the byte shifts I've thought about doing two memory loads.
> Instead of doing the byte shifts I've thought about doing two memory loads. I tried this and it is much slower.
I'm gtg on this PR.
I have documented it here: https://github.com/zlib-ng/zlib-ng/wiki#versioning. Any edits are welcome.
@joshtriplett this should be merged into develop now.
Closing, should be resolved in develop.