Could we agree to disagree and come together on improvements andclean-up?

[Vincent Lefevre]

On 2008-06-02 03:20:13 +1000, Joerg Arndt wrote:
> * NightStrike <nightstrike at gmail.com> [Jun 02. 2008 02:49]:
> > On 6/1/08, Joerg Arndt <arndt at jjj.de> wrote:
> > > Roughly, the code should work on what I call a sane arch/model:
> > > - long is machine word,
> > 
> > If this is what you want, then you are specifically saying "I don't
> > want GMP on Windows -- end of story"
> 
> No!
> Firstly I personally do not care and will not invest any work in that.
> Secondly, I assume (and may be wrong) that having the full performance
> of the 64-bit arch with long=32-bit may need many changes.
> 
> If GMP is mainly needed for gcc then what is the problem with
> suboptimal performance?  I doubt gcc will need much more than 128-bit
> precision!

AFAIK, high precision is at least needed in binary <-> decimal
conversions of (some) floats.

> (At least not a thousand or so).  The time needed inside
> GMP routines should be negligible (again, I could be wrong, but I'd be
> utterly surprised if the performance of GMP would be critical here)!

I don't understand. For the time being, this is not a question of
performance. GCC needs GMP. If GMP doesn't work on some platform,
then GCC won't work either.

-- 
Vincent Lefèvre <vincent at vinc17.org> - Web: <http://www.vinc17.org/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.org/blog/>
Work: CR INRIA - computer arithmetic / Arenaire project (LIP, ENS-Lyon)