The GMP gcd code is not only faster than that in MPIR, it is better.
It's much more cleanly coded. I think MPIR should switch to using the
GMP code.

It's unfortunate that I spent so much time on taking Niels' code and
adding extended GCD and dealing with normalisation and so on, but
didn't have time to really polish it off. Now it is kind of irrelevant
because the GMP code is so well-polished and MPIR can just use it.

The only disadvantage to switching to their code is that there are no
longer two implementations. However, there is still a lot of shared
code either way, as both are developed from Niels' original code and
algorithm. So this "disadvantage" is not so great in my opinion.

Bill.

On 28 September 2011 20:51, Jeff Gilchrist <jeff.gilchr...@gmail.com> wrote:
> On Wed, Sep 28, 2011 at 3:46 PM, Jason <ja...@njkfrudils.plus.com> wrote:
>
>> I found a performance loss of mpir 2.4.0 compared to gmp 4.3.2. when running
>> gmp bench 0.2 gcd examples on a Core 2 Duo T6400, 2Ghz, 4GB Ram Notebook with
>> Ubuntu 11.04, 2.6.38-11-generic-pae i686, gcc 4.5.2.
>
> I'm not sure how GCD code compares in general, but could this be a
> tuning issues with a Core2 laptop?  As in, they should tune the values
> instead of just using the default?
>
> Does the GCD code actually differ a lot between mpir and gmp now? And
> that is with the older gmp 4.3 branch as well.
>
> Jeff.
>
> --
> 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.
>
>

-- 
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