2009/5/3 Marc :
>
> On Apr 30, 5:34 pm, Bill Hart wrote:
>
>> >> * Support for parallel processing (e.g. multithreaded code) and GPU's
>>
>> > This is the part I am curious about: what are your plans there? In
>> > particular, at what level do you place the parallelism? The code is
>> > already t
On Apr 30, 5:34 pm, Bill Hart wrote:
> >> * Support for parallel processing (e.g. multithreaded code) and GPU's
>
> > This is the part I am curious about: what are your plans there? In
> > particular, at what level do you place the parallelism? The code is
> > already thread-safe. Do you want to
2009/5/1 Marc :
>
> Hello,
>
> On Apr 27, 9:47 am, Bill Hart wrote:
>> The current major pushes in MPIR development are:
>>
>> * Faster assembly code on a variety of platforms
>
> (vote for making a mul_1 that uses the new sparc64 VII integer addmul
> instruction, if you can get your hands on the
Hello,
On Apr 27, 9:47 am, Bill Hart wrote:
> The current major pushes in MPIR development are:
>
> * Faster assembly code on a variety of platforms
(vote for making a mul_1 that uses the new sparc64 VII integer addmul
instruction, if you can get your hands on the hardware)
> * Support for par
> My feeling, looking through your code is that:
>
> * It really sits *on top of* the mpz layer
> * It relies on some code for single limbs that is not abstracted
> anywhere in MPIR, but should be
>
> It should:
>
> * be implemented *in* the new mpir_n layer
> * the unsigned long portion should be
The figures you provide are very interesting. Though they illustrate
something interesting. None of the packages which use GMP/MPIR
actually use the code therein for binomial coefficients. They all have
their own code. I suspect this sort of thing happens because a
majority of those packages are b
On ápr. 27, 03:22, Bill Hart wrote:
> Did you receive any feedback
> on the code from them?
As I remember at the same time there were 4 codes from 3 peoples to
speed up the
computation of binomial coefficients. What hasn't been done is to
choose the best one from 5 codes (including original) fo
2009/4/26 gerrob :
>
> Hi!
>
> I've uploaded two codes on the google group, computation of binomial
> coefficients in a much faster way than the currently used in mpir/gmp,
> using binary splitting (the idea is similar if we would use product
> tree). In gmp there are two functions for this: mpz_b