Olivier Giniaux

The `Hasher` trait defines methods to hash specific types. This allows the implementation to circumvent some tricks used when the size is unknown. For this reason, hashing 4 `u32` using...

This should be fairly easy to do. Unfortunately, I don't have much time to do this myself for you in my spare time (kids, you know). I might consider it...

Hello @blyxyas, sorry for the delay. I tried to reproduce back when you opened this issue, but the problem would not arise on my laptop. I am leaving this issue...

Closing stale issue

Hello! > Reading past an array bounds is unsound It may be unsafe, but I don't think it is unsound, because of the check performed. I understand your concern though...

It would be interesting to try! However regarding soundness, from the Miri readme itself: > Moreover, Miri fundamentally cannot tell you whether your code is sound. Soundness is the property...

As said above: > if there is a way to be safer without compromising on performance, we'd gladly take it Otherwise, we'll keep the code as it is. As with...

Thank you for putting in all these efforts. I think your point is clear and I can understand why you feel like it can be "disheartening", however you must understand...

I've renewed the experiment using `simd_masked_load`, but this time I pushed the changes so that I could see what the results are on X86 (because I use an ARM MacBook)...

I created a simple benchmark setup to isolate this out-of-bound optimization alone. You'll find the code in this PR: https://github.com/ogxd/gxhash/pull/118. This benchmark compares: - Copying the exact number of elements...