multi-party-ecdsa
multi-party-ecdsa copied to clipboard
Dev/benchmarks #45
Dummy PR - still may require many different changes, just a draft for @omershlo to review the change to gg18/keygen.rs that I referred to as Black Magik 😄. Also, I found a way to run everything on mobile since there is a branch in rust-crypto where aarch64 was fixed.
@vhnatyk it seems you are using an old version of keygen. Can you use try and use the latest in master please?
@vhnatyk it seems you are using an old version of keygen. Can you use try and use the latest in master please?
Umm, sorry - seems it's latest gg18/keygen.rs, and has only one commit 7483f3880b2995ecf83db25ecafe860f0df29c59. Maybe you refer to 2p keygen bench - lindell_2017/keygen.rs - that one is latest
Oh I see the problem! The keygen for gg18 in benches that you are referring to is not up to date - compare it with the test in master: https://github.com/KZen-networks/multi-party-ecdsa/blob/master/src/protocols/multi_party_ecdsa/gg_2018/test.rs#L56
If you want to get rid of the black magic - copy the code from test.rs instead of what we have now in benches. let me know if you prefer me to do it [its suppose to be straight forward]
It will also allow you to run with different t,n
parameters easily.
If it's just simple overwriting update to gg18\keygen.rs - yes pls update, otherwise if some merging etc - probably I can do it and you'll review. Cool - let's get rid of Black Magik ! 🤘 - I will rerun benches with latest
I pushed the new GG18 keygen benchmarks to master
.
- I changed Cargo.toml to point benchmark to this file
- write now I bench two random cases (t,n = 1,2 and 2,3), feel free to expend and add stuff according to #45