2 ms·
Marz's taylor series implementation as-is is significantly faster (~60% the runtime) than the glibc implementation at 6 terms on My Machine(tm), rather than abo
by aidenn0 3y ago
Marz's taylor series implementation as-is is significantly faster (~60% the runtime) than the glibc implementation at 6 terms on My Machine(tm), rather than about the same as in TFA. I haven't compared to LERP lookup table yet though.