Open DanieleDiBenedetto opened 4 years ago
Originally the code was using projective formulas, and I actually think I might switch to them (after some benchmarks) but I wanted to have an implementation with Jacobian to benchmark against after other optimizations have been written, that way I know it's worth it. (#2)
The choice of representation here doesn't matter though. The API does not provide access to the raw coordinates, so this can be switched later without breaking anything.
Hi, shouldn't the zero point in SW Jacobian be any point of the form (t^2:t^3:0) with t <> 0 (conventionally represented with (1:1:0)), instead of (0:1:0), which is actually the zero point in SWProjective ? https://github.com/zkcrypto/bls12_381/blob/e32494e72011b9811bd7f5c34af5ff6785cc75db/src/g1.rs#L516