[MPFR] Re: Speed of GMP relative to MPFR for quad precision calculations

Vincent Lefevre vincent at vinc17.net
Tue Nov 20 00:40:26 CET 2012


On 2012-11-19 16:23:17 -0700, Jerry wrote:
> Is double-double part defined in C?

It is part of the PowerPC ABI, where long double = double-double
arithmetic. Otherwise, it needs to be re-implemented by the user,
based on IEEE 754 (this won't work on Linux/i386 machines due to
the use of extended precision; you need a processor with SSE2
support).

> Is this a way of getting (near-)quad precision using double
> precision floating point hardware?

One can say that. You get something like a 106-bit precision (it
can be more in specific cases), instead of 113.

-- 
Vincent Lefèvre <vincent at vinc17.net> - Web: <http://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)


More information about the gmp-discuss mailing list