Daniele Panozzo 35cd7d4dc2 optimized the precomputation of sparse_fast, now it is comparable to sparse (2-3x slower) %!s(int64=8) %!d(string=hai) anos
..
igl 35cd7d4dc2 optimized the precomputation of sparse_fast, now it is comparable to sparse (2-3x slower) %!s(int64=8) %!d(string=hai) anos