Re: TomsFastMath v0.10 released



Wolfgang Ehrhardt wrote:
Is there any specific reason why TFM uses a simple Euclidian gcd
(without special fast mod implementation) and not at least the binary
gcd from standard LTM (not to mention faster algorithms)?

Because I know everything, I'm the bestest, lalalala.

No, because like the LTM case it wasn't a focus at the time. Believe
it or not 99% of your cpu time is NOT in invmod. So while I agree I
should probably invest some time in speeding it up, it's wasn't a
priority.

Also, I just started on the LT projects again [was working on that book
that I keep hearing so much about...] so the TFM project has been kinda
stagnating.

If you want to contribute a faster invmod then by all means do so and
I'll include it.

tom

.