Alexander Kruppa
72fe382864
sqr_basecase for Skylake from GMP 6.1.1
2016-12-20 06:32:39 +01:00
Alexander Kruppa
5ae24aef66
mul_basecase for Skylake from GMP 6.1.1
2016-12-20 05:15:54 +01:00
Alexander Kruppa
f28068172d
add_n and sub_n
...
1c/l most of the time, but sometimes gets into a bad "mode" where performance
degrades to up to 1.2c/l
2016-12-20 05:14:15 +01:00
Alexander Kruppa
cfc589609e
Move to haswell/
...
This sumdiff_n is much slower on Haswell (2.6c/l) than on Skylake (2c/l)
but it still provides a ~3% speed up for a 1M limb FFT compared to having
no sumdiff_n at all.
2016-12-08 16:23:48 +01:00
Alexander Kruppa
4459641bad
sumdiff_n optimized for Skylake
...
2c/l
2016-12-05 16:40:57 +01:00
