On Friday 09 December 2011 03:57:58 David Cleaver wrote:
> Hello all,
> 
> I've been using 2.4.0 successfully for a while now.  I've downloaded
> 2.5.0-rc1 and was able to compile it and use it to build GMP-ECM with no
> problems (it even makes GMP-ECM a tiny bit faster).  However, I wanted to
> run the tuneup program to try to optimize mpir for my system.  I read in
> the manual that I should be able to:
> cd tune
> make tuneup
> ./tuneup
> However, "make tuneup" gives a lot strange error messages.  For some
> strange reason, in order to build tuneup on my computer, I need to do:
> cd tune
> make tuneup.exe
> ./tuneup
> This creates tuneup.exe just fine.
> 
> Then, when I run tuneup.exe, the program seemingly hangs at:
> #define HGCD_THRESHOLD                   30
> #define GCD_DC_THRESHOLD
> It turns out that it is not hanging, it gets to line 2374 of speed.h and
> (somehow) a very large number is copied from s->reps into i.  Then, since i
> is so large, the following "do" loop just keeps going and going.
> 
> It took a while to track down the problem to this part of the code, and I'm
> not quite sure:
> 1) how s->reps gets set
> 2) why s->reps is so large during this function and not any others
> 
> Has anyone come across this problem before?  Is there a simple solution? 
> Would anyone need more information to help me figure this out?  If so,
> what would you like to know?
> 
> My environment is as follows:
> Windows XP Pro x64 SP2, Msys shell, Mingw64 compiler, MPIR 2.4.0 and/or
> MPIR-2.5.0-rc1.
> 
> Any pointers or suggestions on how to proceed would be much appreciated. 
> Thanks for your time.
> 
> -David C.

Hi

Tuning under windows is very problematic and the difference between windows and 
unix tuning values is minimal . Hopefully in the next MPIR release the tuning 
will work correctly under all OS'es . And note the tuning for the above part 
you 
mention is slow on all systems , perhaps with windows slightly more jitter in 
the timings would make it very slow.

Jason

-- 
You received this message because you are subscribed to the Google Groups 
"mpir-devel" group.
To post to this group, send email to mpir-devel@googlegroups.com.
To unsubscribe from this group, send email to 
mpir-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/mpir-devel?hl=en.

Reply via email to