Aleksey Vaneev
Aleksey Vaneev
With 32-bit hashes I've meant there will be collisions with 480k entries, this will reduce performance. (In SMHasher3 there are 790k entries in hash-map test). I've noticed SMHasher3 and likely...
In SMHasher3, if you cast a5hash32() result to `uint64_t` and use it just for hash-map test, the performance degradation goes away.
If you are going to fix the `a5hash128` call, fetch a newer version - 5.12
Posted a5hash v5.13 - some minor rearrangements, but this may change the code size and timings a bit. They are better on my machines and my own benchmarks, but in...
It would be great if you tested and posted a5hash128 5.16. I wonder how it works on Zen+. On Zen2 it's as fast as wyhash, and much faster on Zen5....
It's also worth updating speeds and code size of a5hash() - its code was rearranged since last time you tested it. In my tests, the latest version is slightly better,...
It's interesting lzav -2 is on par or outperforms zlib -9 on larger files. lzav is pure LZ77, and its compression algorithms were tuned for speed, so its stream format...
Well, as an author of lzav I generally rarely witnessed -2 being faster than -1 at decompression. It may be faster maybe by 2%, and not by 10%. The contradiction...
Well, probably I'm incorrect and there can be edge cases.
What matters it's both the number of matches and average match length. Many short matches yields slower decompression - that's what I usually observe with silesia dataset. Thanks for the...