search

Aatch / ramp

RAMP - Rust Arithmetic in Multiple Precision
Apache License 2.0
263 stars 38 forks source link

AVX2 #111

Open burdges opened 5 years ago

burdges commented 5 years ago

YMP claims an impressive 2-3x speedup over GMP by using AVX2. Any opinions on: If what YMP does is sensible. What AVX2 support might look like? What internal algorithms it'd improve? How much work it requires?

rozbb commented 5 years ago

Hey, sorry for the delay on this. The original author of this crate isn't really around to do more dev work, and I don't think I have the expertise to do anything beyond basic bugfixes and small API changes at this point. I would normally say that I'd be happy to accept a PR, but I think it would be irresponsible of me to do that for something as complicated as an AVX2 bigint implementation.

Thank you for bringing this up. This is a good open question, so I'll leave the issue open.

burdges commented 5 years ago

Cool, thanks! We'll select MPIR over GMP or ramp: https://github.com/poanetwork/vdf/issues/13

If anyone wants funding to work on ramp, then maybe we can help arrange something, but not sure it's such a high priority for us now anyways. I care much more about MLS btw ;)