Closed fionser closed 6 years ago
Hi, I found that the 'ntt_new' method in 'tests/ntt_pefs.cpp' is 2x faster than the ntt_loop in the NFLib, even I used AVX and SSE opts.
Can any one tell me why this performance gap happen ?
Hi, I found that the 'ntt_new' method in 'tests/ntt_pefs.cpp' is 2x faster than the ntt_loop in the NFLib, even I used AVX and SSE opts.
Can any one tell me why this performance gap happen ?