[Wannier] Possible bug in postw90/dos.F90

[Giovanni Pizzi]

Dear Brad,

first of all thanks a lot for the report.
Unfortunately, is is not easy to understand where the problem lies with the information that you provide.

The warning messages should not be harmful (they just mean that some arrays needed to be copied in a temporary location because the array passed to the caller was not contiguous; this can result in a performance loss, but should not generate any problem).

Can you identify more specifically where the code crashes, e.g. by looking at the last line printed by the code, or using a debugger as gdb after compiling with the the -g option and then watching at which point of the code the crash occurs? In particular, it would be useful to know if the bug really happens inside the dos subroutines (or even in which of the called subroutines), or outside of them.

Moreover, could you also provide the input file, if possible, to understand which part of the code is used and could trigger the problem?

Best,
Giovanni


--
Giovanni Pizzi
Post-doctoral Research Scientist
EPFL STI IMX THEOS
MXC 340 (Bâtiment MXC)
Station 12
CH-1015 Lausanne (Switzerland)
Phone: +41 21 69 31124




On 30 May 2014, at 23:00, Brad Barker wrote:

Hello Wannier Mailing Group,

I am a new user of Wannier90. I have compiled Wannier90-2.0.0 on the NERSC machine Edison, which uses an Intel compiler.

The below happens when I run postw90 to calculate Projected DOS calculations on a subset of Wannier functions.

Using the default optimization flag in the make.sys file

FCOPTS=-O2
LDOPTS=-O2

postw90 produces the error message

forrtl: severe (174): SIGSEGV, segmentation fault occurred

and the execution is terminated.

Modifying the make.sys file to include more debugging flags

FCOPTS=-O2 -g -traceback -check all -fp-stack-check
LDOPTS=-O2 -g -traceback -check all -fp-stack-check

produces the messages

forrtl: warning (402): fort: (1): In call to KMESH_GET_BVECTORS, an array temporary was created for argument #4

forrtl: warning (402): fort: (1): In call to I/O Write routine, an array temporary was created for argument #2

for each processor, and now postw90 executes successfully.

I am unsure if the developers and the community are aware of this problem or possible bug fixes in postw90/dos.F90.

Thank you,
Bradford Barker
University of California, Berkeley
_______________________________________________
Wannier mailing list
Wannier at quantum-espresso.org<mailto:Wannier at quantum-espresso.org>
http://www.democritos.it/mailman/listinfo/wannier

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/wannier/attachments/20140530/c058ede9/attachment.html>