foxmuldrster at yahoo.com foxmuldrster at yahoo.com
Fri Dec 3 03:47:17 CET 2010

I've been looking at this problem today. I am going to need far better  
developer tools than are currently available to do this. I'll begin there.

- Rick

-----Original message-----
From: Paul Zimmermann <Paul.Zimmermann at loria.fr>
To: "Rick C. Hodgin" <foxmuldrster at yahoo.com>
Cc: gmp-devel at gmplib.org, mpfr at loria.fr
Sent: Wed, Dec 1, 2010 12:49:23 GMT+00:00
Subject: Re: [MPFR] GMP+MPFR to GPU?


> How well suited to parallel / vector processing are GMP's and MPFR's
> algorithms?  If, for example, we setup N identical operations for
> parallel computing (adds, multiplies, cosines, whatever), how would the
> existing design handle the vectorization?

currently there is no vectorization in MPFR (and in GMP also I guess).
Thus the "vectorization" should be done at the thread level.


More information about the gmp-discuss mailing list