Port blake2 on coresimd

[newpavlov]

With stabilization of SIMD intrinsics in Rust 1.27 we can remove simd feature from the blake2 crate.

[tomtau]

btw, here's a bounty on this issue: https://www.bountysource.com/issues/59948303-port-blake2-on-coresimd

[kazcw]

How would this work for rustc 1.21..=1.26?

[newpavlov]

For v0.8 it will have to be feature-gated behind some feature (it can be the same simd feature), but I am thinking about migrating to Rust 2018 for v0.9.

[DonaldTsang]

What about BLAKE3?

[tarcieri]

I’d be very much into us having a high-performance BLAKE3 implementation but am not aware of the state-of-the-art in Rust BLAKE3 libraries /cc @zooko

[tarcieri]

See BLAKE3 #103

[pczarn]

I started working on the issue.

[tarcieri]

@pczarn it'd be really good to get #228 landed before we try to extend the current implementation

[pczarn]

@tarcieri Okay, so you are close to getting a new implementation, but a migration from core::arch x86/x86_64 intrinsics to a portable core::simd will remain to be done?

[pczarn]

Looks like coresimd is some internal utility, and not core::simd, so:

• could you point me to the code for coresimd?
• would it be better to give it a name that cannot be confused with core's simd?
• are you looking into using core's simd once it gets stabilized?

[tarcieri]

I think this ticket is actually referring to core::arch SIMD intrinsics, though adding support for core::simd would be nice

[pczarn]

Ok, please update me when progress with #228 happens