[Pw_forum] array bound check

[Manoj Srivastava]

Dear Dr. Giannozzi,Dr. Baroni and Dr. Paulatto, 
 Thank you very much for your help. I finally have done the array bound
check in the desired subroutine. But the array that goes out of bound, at
least that is the message I get, I dont think it could go out of bound.
Could this be problem of the compiler?  I have noticed that sometimes when
I run a job, it gives me some random error, but when i run the same job
again, it runs without any complain! I am using intel ifort compiler on
x86_64 architecture machine.

Regards, 
Manoj Srivastava

 
On Thu, 25 Sep 2008,
Paolo Giannozzi wrote:

> 
> On Sep 25, 2008, at 6:41 , Manoj Srivastava wrote:
> 
> > I guess I will try the CVS version and see what i get.
> 
> I guess you should rather see what happens at the line of the
> code where the error happens. If you did things properly (likely),
> and if the compiler work properly (less likely), you should have
> gotten a message saying something very self-explanatory like
> 
> Subscript #1 of the array R has value 172 which is greater than the  
> upper bound of 171
> 
> in addition to a traceback with the line number where the offending
> command is called.
> 
> Sometimes such messages are bogus, i.e. they do not signal a true
> out of bound array but usage of obsolete syntax that prevents the
> compiler to know about the true size of arrays
> 
> > By the way, is there a way to do array bound check in a particular
> > subroutine I am interested rather than doing the whole code?
> 
> there is the safe and slow way: compile normally the entire code,
> change the flags in make.sys to switch bound check on, remove
> the object (*.o) corresponding to the routine you want to check,
> "make" again
> 
> Paolo
> ---
> Paolo Giannozzi, Dept of Physics, University of Udine
> via delle Scienze 208, 33100 Udine, Italy
> Phone +39-0432-558216, fax +39-0432-558222
> 
> 
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>