2 ms·
I assume double-double would be the emulated __float128 type in GCC (or hand written equivalent)? Can GCC vectorize it or it needs to be done by hand?
by gpderetta 5y ago
I assume double-double would be the emulated __float128 type in GCC (or hand written equivalent)? Can GCC vectorize it or it needs to be done by hand?