Re: [Gretl-users] CVS bug-fixing report
Allin Cottrell wrote:
Thanks very much Marcin.
sorry...
It's a bit hard to read this trace, since valgrind ends up
aborting
when it runs into a BLAS 3dnow! instruction.
I know that there are problems with
AMD and linear algebra libraries; I
use ATLAS 3DNow!, and that may be problem; I don't remember where, but
I've read, that in fact many applications runs wrong with these
libraries (3DNow is not IEEE-compliant), but I've never experienced
problems;
I'll try the same script on my note, there I have Intel with SSE;
Marcin
I wonder if this _might_ be a problem with the BLAS implementation
you're using. There is some discussion of possible Debian LAPACK/BLAS
flakiness in relation to R:
http://bugs.r-project.org/cgi-bin/R/System-specific?id=8041;user=guest
and also Octave:
http://www.octave.org/mailing-lists/bug-octave/2005/1120
Allin.
_______________________________________________
Gretl-users mailing list
Gretl-users(a)ricardo.ecn.wfu.edu
http://ricardo.ecn.wfu.edu/mailman/listinfo/gretl-users
--
Marcin Błażejowski
http://www.wrzosy.nsb.pl/~marcin/
GG# 203127