The underlying k256 backend has Projective + Affine addition (a little faster than Projective + Projective). We had at some point stopped using it (or maybe we never were!). This reintroduces it while also cleaning up the backend API a bit.
The actual perf improvement is ~310ns instead of ~335ns per addition.
The underlying k256 backend has Projective + Affine addition (a little faster than Projective + Projective). We had at some point stopped using it (or maybe we never were!). This reintroduces it while also cleaning up the backend API a bit.
The actual perf improvement is ~310ns instead of ~335ns per addition.