[Wannier] problem with num_exclude_bands while using w90chk2chk.x

[Giovanni Pizzi]

Dear Krzysztof,

thanks a lot for the report.

Indeed, as you correctly spotted, there was a missing allocation in the specific case with num_exclude_bands > 0.
Actually, the same problem was also present in the routine reading from a formatted file, needed for the conversion in the opposite format.

I committed the correction to the Wannier CVS and this will be distributed with the next Wannier90 release.
Would you be so kind to check if the attached w90chk2chk.F90 (which is the fixed version) works for you, when converting in both directions (e.g. you can convert to formatted, then back to unformatted, and see if you are e.g. still able to restart from the chk file generated with the roundtrip and obtain the same results).

Thank you in advance,
Giovanni Pizzi


--
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 28 Feb 2015, at 12:36, Dymkowski Krzysztof Pawel wrote:

Dear All,
I was recently trying to convert my seedname.chk to seedname.chk.fmt using utility tool: w90chk2chk.x. I was able to run tests provided without any errors. I also was able to produce seedname.chk.fmt for both test1 and test2.

Now  I am trying to  do NiO benchmark (paramagnetic calculations) and do later on to use of this for  DMFT calculations.  When runnig w90chk2chk.x I got segmentation fault. I have added two lines to w90chk2chk.f90:

allocate(exclude_bands(num_exclude_bands),stat=ierr) ! ****KD*****
if (ierr/=0) call io_error('Error allocating exclude_bands in param_read') !*****KD*****

And it seems to work for my benchmark. In the attachment FILES I am sending you the main inptu/output files from the run with modified w90chk2chk.f90.

Without modification I get the following error while runnig w90chk2chk.x:

forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image              PC                Routine            Line        Source
w90chk2chk.x       00000000005C3FE1  Unknown               Unknown  Unknown
w90chk2chk.x       00000000005C2737  Unknown               Unknown  Unknown
w90chk2chk.x       000000000056FE64  Unknown               Unknown  Unknown
w90chk2chk.x       000000000056FC76  Unknown               Unknown  Unknown
w90chk2chk.x       000000000052892F  Unknown               Unknown  Unknown
w90chk2chk.x       000000000052C92D  Unknown               Unknown  Unknown
libpthread.so.0    00007FABA32A40A0  Unknown               Unknown  Unknown
w90chk2chk.x       00000000005D2D26  Unknown               Unknown  Unknown
w90chk2chk.x       0000000000548D2D  Unknown               Unknown  Unknown
w90chk2chk.x       0000000000546026  Unknown               Unknown  Unknown
w90chk2chk.x       0000000000405A57  w90_conv_mp_conv_         115  w90chk2chk.F90
w90chk2chk.x       00000000004040B8  MAIN__                    593  w90chk2chk.F90
w90chk2chk.x       0000000000403AFE  Unknown               Unknown  Unknown
libc.so.6          00007FABA2AA2EAD  Unknown               Unknown  Unknown
w90chk2chk.x       00000000004039E9  Unknown               Unknown  Unknown

I have attached log file from w90chk2chk  (w90chk2chk_vannila.log‎) from run without any modifications. It stops at work after reading bands.


I have compiled wannier90 using the following settings:
F90 = /opt/intel/bin/ifort
#COMMS=
MPIF90=/opt/intel/impi/5.0.3.048/bin64/mpif90
FCOPTS=-O2 -heap-arrays  -g -traceback
LDOPTS=-O2 -g -traceback

LIBDIR = /opt/intel/mkl/lib/intel64
LIBS   =  -L$(LIBDIR) -lmkl_core -lmkl_intel_lp64 -lmkl_sequential -lpthread

Is there a problem with my compilation or have  I  found a bug?

Regards,
Krzysztof Dymkowski



<FILES.tar.bz2><w90chk2chk_vannila.log>_______________________________________________
Wannier mailing list
Wannier at quantum-espresso.org<mailto:Wannier at quantum-espresso.org>
http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/wannier/attachments/20150304/0c2f1b76/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: w90chk2chk.F90
Type: application/octet-stream
Size: 24590 bytes
Desc: w90chk2chk.F90
URL: <http://lists.quantum-espresso.org/pipermail/wannier/attachments/20150304/0c2f1b76/attachment.obj>