Portable and compile ver of GMP
Sam Rawlins
sam.rawlins at gmail.com
Wed Dec 19 01:53:20 CET 2012
On Tue, Dec 18, 2012 at 2:52 PM, <bodrato at mail.dm.unipi.it> wrote:
>
>
> The paper you refer to reads: "We considered two versions of GMP. The
> first was the compiled version of GMP which is highly optimized for the
> underlying microprocessor. The other was the highly portable version of
> GMP". Right?
>
Are there any links to this paper? I think we're all in the dark unless we
know what is being talking about.
>
> What do the authors mean with "portable version"? We should ask them.
>
Yes, I think the authors are the only people qualified to answer that
question.
> I can guess they mean the same GMP library, compiled with the
> --disable-assembly (was --host=none...) configure option.
> In both cases, (with or without assembly portions) the library uses
> different algorithms "depending on the data size", for some tasks, e.g.
> multiplication. The goal is, of course, to use the faster algorithm in
> each range.
>
> Regards,
>
> --
> http://bodrato.it/
>
> _______________________________________________
> gmp-discuss mailing list
> gmp-discuss at gmplib.org
> https://gmplib.org/mailman/listinfo/gmp-discuss
>
--
Sam Rawlins
More information about the gmp-discuss
mailing list