[Beowulf] Re: progress on ifort and gfortran issue

f.costen@cs.man.ac.uk fumie.costen at manchester.ac.uk
Thu Apr 9 12:06:03 EDT 2009


Dear All,
Finally this issue is solved
by one of my students, Philip Clapham .

Assume we have an array  which is
declared as an  allocatable array.
When this array is used without explicit allocation
of memory by mistake,
ifort and gfortran  implicitly allocate
the memory in a different way.

ifort gives the size of 0
and
gfortran gives the size of 1

This was very difficult to track using the
debuggers and we actually put the manual check
line by line...

Anyway thank you very much indeed for
all of your help on this issue.

Have a lovely and long Easter weekend :-)
Best wishes, Fumie


f.costen at cs.man.ac.uk wrote:
> Dear Mark, Dmitri, Xavier, Dave, and Tom,
> thank you very much for your suggestions.
> First of all, I tried the compilation
> with -fbounds-check option and I did not get any error
> message.
> When I checked the site http://gcc.gnu.org/wiki/GFortran
> our gfortran which shows:
> fumie at e3sfc32:~/ngs-fdcn-code> gfortran --version
> GNU Fortran (GCC) 4.4.0 20081011 (experimental) [trunk revision 141054]
> Copyright (C) 2008 Free Software Foundation, Inc.
> seems to be the latest now.
> 
> There is no file to read in as this is the
> very primitive code for the new algorithm test,
>  not the production code(=reference code).
> 
> We are going to try to learn  gdb so that
> we can understand where the segfault comes from.
> 
> I will keep you updated.
> Thank you very much for your help
> Best wishes, Fumie
> 
_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.



More information about the Beowulf mailing list