--- Jud McCranie <j.mccranie@...
> At 09:42 AM 6/3/2006, hecht wrote:
> >I can factor numbers in the 12 digit range pretty quickly now, but it still
> >took over a minute to find a factor of 10^15 + 3.
> >Why so long? It takes me 24 ms (Pollard Rho, HIT library, PIII 700 MHz)
> For numbers in that range (fewer than 64 bits), I wouldn't use HIT or any
> other such package. Just use the native data types. Also, Shank's method
> is probably better for that application than Pollard's method. My program
> on a 2.4 GHz factors 10^15+3 in 0.01 second.
Now I've decomissioned my PPro/200, I don't have a machine slow
enough to measure my native double reimplimantation of Lenstra's
(LIP) version of Shanks' SQUFOF on that number! (And I gave away
my ARM dev board yesterday - that would have been really slow.)
However, I found two Sparcs, with their notoriously underpowered FPUs,
in America, and indeed they are slow enough for it to be measurable:
SunBlade100 - 0.007s
SunUltra5 - 0.010s
(no idea what clock speed they may be).
Paul - may I post the code? It's Lenstra's copyright, but you seem
to be the guardian of it presently?
() ASCII ribbon campaign () Hopeless ribbon campaign
/\ against HTML mail /\ against gratuitous bloodshed
[stolen with permission from Daniel B. Cristofani]
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around