Re:Improving the fast inverse square root (2010)
by robin_reala on Monday July 09, 2018 @12:47AM
(#17488031)
How does that perform in comparison?
[ Parent
| on HN ]
Re:Improving the fast inverse square root (2010)
by brandmeyer on Monday July 09, 2018 @01:01AM
(#17488072)
rsqrt{p,s}s has guaranteed relative error <= 1.5 * 2^-12, or about 3.6e-4. According to Agner Fog, it typically executes in one cycle. I would assume that the AVX512 versions are similar.