[Wien] Bug in seclr4.F (probably dsyxev4.f or dstebz2.f) for MKL_NUM_THREADS>1
Laurence Marks
L-marks at northwestern.edu
Fri Apr 11 13:48:44 CEST 2008
Thinking about it, I suspect that with multiple-threads (which is
going to be with us for the new quad-cores) workspace requirements
will be larger by a multiple of the number of threads and this will
cause problems. While this maybe the first definitive case, I would
not be surprised if similar issues arise in other parts of the code. I
don't think the is a bug in mkl10.x, rather that there will need to be
some changes in the code due to how threads are handled.
On Fri, Apr 11, 2008 at 1:34 AM, Peter Blaha
<pblaha at theochem.tuwien.ac.at> wrote:
> Many people reported problems with mkl10.X (which are not present in mkl9.x
> and presumable there's a bug in the mkl10.x
>
> Nevertheless, and can print the relevant numbers in the call of
>
> CALL DLATRD( UPLO, N-I+1, NB, A( I, I ), LDA, E( I ), &
> TAU( I ), WORK, LDWORK )
>
> ....
>
> Laurence Marks schrieb:
>
> > The array IWORK is getting corrupted, and the deallocation on line 367
> > is failing with a SIGSEV.
> >
> > Compiler: ifort 10.1.015, mkl 10.0.2.018
> > libpthread-2.5.so glibc-2.5.s0 libgcc_s-4.1.1-20070105
> >
>
> --
>
> P.Blaha
> --------------------------------------------------------------------------
> Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
> Phone: +43-1-58801-15671 FAX: +43-1-58801-15698
> Email: blaha at theochem.tuwien.ac.at WWW: http://info.tuwien.ac.at/theochem/
> --------------------------------------------------------------------------
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>
--
Laurence Marks
Department of Materials Science and Engineering
MSE Rm 2036 Cook Hall
2220 N Campus Drive
Northwestern University
Evanston, IL 60208, USA
Tel: (847) 491-3996 Fax: (847) 491-7820
email: L-marks at northwestern dot edu
Web: www.numis.northwestern.edu
Commission on Electron Diffraction of IUCR
www.numis.northwestern.edu/IUCR_CED
More information about the Wien
mailing list