Aleksandar Donev wrote:
>
> Hello,
>
> I am observing something I have not seen before: a massive loss of precision
> with one compiler (out of 3). I have some calculations which have certain
> mathematical properties, such as the norm of a quaternion being preserved at
> 1. With two compilers this is correct within numerical (double) precision,
> 1E-15 or alike. But with one compiler I am getting much larger discrepancies,
> order 1E-13 to 1E-12, and these eventually lead to my whole simulation
> breaking down. The calculations involved are many small matrix operations
> (inverses, multiplications), so it is quite imaginable there would be
> significant roundoff.
>
> What could this be related to? Is it optimization, possible mismatch between
> compiler and platform (the compiler is older then the processor), or should I
> be checking my codes?
>
> Any hints from people who have experienced this would help.
> Thanks,
> Aleksandar
> --
> __________________________________
> Aleksandar Donev
> Complex Materials Theory Group (http://cherrypit.princeton.edu/)
> Princeton Materials Institute &
> Program in Applied and Computational Mathematics
> @ Princeton University
> Address:
> 419 Bowen Hall, 70 Prospect Avenue
> Princeton University
> Princeton, NJ 08540-5211
> E-mail: [log in to unmask]
> WWW: http://atom.princeton.edu/donev
> Phone: (609) 258-2775
> Fax: (609) 258-1177
> __________________________________
Hi,
very recently I had a similar problem of loss of accuracy. It turned out that
in operations A**B with both A and B of quadruple precision the result turned
out to be returned with precision of 8 bytes only. Clearly this is a compiler
bug and I informed the vendor on this bug.
Regards,
Volker
--
****************************************
* *
* Dr. Volker Typke *
* University of Ulm *
* Communication & Information Center *
* Albert-Einstein-Allee 11 *
* D-89069 Ulm, Germany *
* Phone: +731-50-22473 *
* Fax: +731-50-22471 *
* e-mail: [log in to unmask] *<= ATTENTION: slightly changed address !
* WWW: http://www.uni-ulm.de/~typke *
* *
****************************************
Ein Spezialist ist ein Mensch,
der von immer weniger immer mehr versteht,
bis er von nichts alles versteht.
(Kanon von Herbert Beuerle nach einem Text aus Maria Laach)
|