-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
Hi Roberto,
in my project I would need this for the polytope implementation. Anyways, I can have a look at the checked source files and see, what I can do for mpfr support.
Best regards, Stefan
Am 26/07/14 10:34, schrieb Roberto Bagnara:
Hi Stefan.
On 07/25/14 10:21, Stefan Schupp wrote:
my idea is as follows: I have a lightweight wrapper for different floating point implementations (e.g. mpfr, gmp, native float types) and rational types (not yet implemented). I want to use this wrapper as a number type for the PPL, for example to be able to make use of the multiprecision features of mpfr.
Just to make sure, you want to use those for boxes, bounded-difference and octagonal shapes approximations, right? Because it is not possible to use floating-point coefficients for the other approximations.
Moreover, for boxes, bounded-difference and octagonal shapes, GMP's rationals and native floating-point types are already supported. Furthermore, as far as I now, GMP's floating-point numbers are deprecated in favor of MPFR's ones.
So the only thing missing is support for MPFR. Are you willing to work on that? The first thing to do is to familiarize with Checked* and checked* source files. Kind regards,
Roberto
Am 25/07/14 09:42, schrieb Roberto Bagnara:
On 07/24/14 15:57, Stefan Schupp wrote:
How exactly to I extend the PPL (if possible) to my own number types?
Hello Stefan,
extending the PPL to use other numeric types can be done, but not via the ordinary user interfaces. Can you give us an idea about what you are trying to accomplish? Kind regards,
Roberto
_______________________________________________ PPL-devel mailing list PPL-devel@cs.unipr.it http://www.cs.unipr.it/mailman/listinfo/ppl-devel
- -- Stefan Schupp M.Sc. RWTH Aachen University Computer Science Department, Informatik 2 D-52056 Aachen, Germany http://www-i2.informatik.rwth-aachen.de/i2/schupp/ Tel.: +49 241 80 21243