|Bill Allombert on Fri, 07 Nov 2003 12:29:09 +0100|
[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]
|Re: Experimenting with the Alpha build.|
On Fri, Nov 07, 2003 at 12:09:56PM +0100, Karim BELABAS wrote: > On Thu, 6 Nov 2003, Phil Carmody wrote: > > --- Karim BELABAS <Karim.Belabas@math.u-psud.fr> wrote: > >> On Thu, 6 Nov 2003, Phil Carmody wrote: > >>> I was wondering if -ffast-math would do to Pari/GP/GMP on the alpha. > >>> The speed-ups weren't significant on the whole (I should have tried > >>> -mno-ieee too perhaps) > >> > >> This is expected: there's almost no floating point code in PARI (everything > >> is emulated using 'unsigned long' arithmetic). Might be considered a > >> shortcoming. > >> > >> There are lots of variables of type 'double', but they are used in parts of > >> the code which is negligible timewise. And never where stability / accuracy > >> is important. I.e --fast-math should be a no-op with respect to performances > >> or results. > > > > I also recompiled GMP with -ffast-math too for the test. In theory GMP might > > perform some FP maths (FFT multiplication? or is it NTT?). > > NTT from a quick glance at the code. I don't think GMP uses doubles much, > either. mp_limb_t is always a variant of some unsigned integer (int, long, > long long). What is NTT ? GMP FFT use modular arithmetic, not double. Cheers, Bill.