search

briansmith / crypto-bench

Benchmarks for crypto libraries (in Rust, or with Rust bindings)
70 stars 11 forks source link

Use `AsMut<[u8]>` implementation instead of slices #2

Closed hauleth closed 8 years ago

hauleth commented 8 years ago

This speeds up Octavo results by order of magnitude. Unfortunately it is still impossible to use in case of SHA512 as there is no implementation of AsMut<[u8]> for &mut [_; 64].

briansmith commented 8 years ago

Wow, really! I didn't expect this to have any impact on the performance.

I'll return to this later. I want to investigate the perf difference myself and probably file a bug against rustc or the Rust library.

hauleth commented 8 years ago

It is probably due unneded bounds checks but I haven't investigate it further.

Łukasz Jan Niemier

Dnia 26 paź 2015 o godz. 18:47 Brian Smith notifications@github.com napisał(a):

Wow, really! I didn't expect this to have any impact on the performance.

I'll return to this later. I want to investigate the perf difference myself and probably file a bug against rustc or the Rust library.

— Reply to this email directly or view it on GitHub.

hauleth commented 8 years ago

No longer relevant as on current nightly there is no difference. Closing.