I've merged the code
Thanks Robert :)

I made a few minor changes to suit MPIR better , changed malloc/free
to the standard mpir ones
__GMP_ALLOCATE_FUNC_.../__GMP_FREE_FUNC_... and replaced the floating
point sqrt with a purely integer one. I added a threshold and
tuning ,to switch between the old and new factorial code. I made no
attempt to change/optimize anything else.
Your new code cuts in at about n=8192 on a nehalem

Thanks
Jason


On 17 July, 18:16, Bill Hart <[email protected]> wrote:
> I've added a ticket regarding merging this code:
>
> http://trac.mpir.org/mpir_trac/ticket/226
>
> I've been unable to find the time myself. Volunteers are welcome.
>
> Bill.
>
> On 20 June, 12:03, William Stein <[email protected]> wrote:
>
> > On Sat, Jun 20, 2009 at 1:00 PM, Bill Hart<[email protected]> 
> > wrote:
>
> > > I think this will definitely be suitable for inclusion in MPIR.
>
> > > I haven't fully made up my mind, but I *think* I am going to spend the
> > > next two days merging all Robert's contributions. MPIR-1.2.2 will
> > > basically be for that purpose, I think.
>
> > Awesome!  Since it doesn't fit so naturally into Sage, but I know a
> > lot of people benchmark
> > Sage by computing n factorial, for some reason, so it's good if
> > factorial is made faster.  Also,
> > of course it comes up in various calculations and having it in MPIR is
> > the best way to make
> > sure the faster version is used in all components of sage (that use MPIR).
>
> > William
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"mpir-devel" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/mpir-devel?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to