I am libsecp256k1’s codebase, for studying causes (doubt I can contribute with something helpful there). Whereas wanting on the area implementation, the implementation of secp256k1_fe_mul_inner in each 10×26 and 5×52 is a bit of unusual to me. I do know some multiplication algorithms, like Karatsuba and Toom’s algorithms, however this one does not appear to be one in every of then, at the very least I could not relate.
Trying on the supply and git historical past, I solely see a reference to Peter Dettman, however I used to be unable to seek out any paper authored by him. Is that this algorithm impressed by one public paper? Any reference can be useful, as I am unable to observe motive concerning the code itself. I perceive what it’s doing, however I do not know precisely how.