GW calculation using 5.0 fails with segmentation fault (signal 11)

You can find here problems arising when using old releases of Yambo (< 5.0). Issues as parallelization strategy, performance issues and other technical aspects.

Moderators: Davide Sangalli, andrea.ferretti, myrta gruning, andrea marini, Daniele Varsano, Conor Hogan

Bramhachari Khamari
Posts: 96
Joined: Wed Oct 23, 2019 5:52 am

Re: GW calculation using 5.0 fails with segmentation fault (signal 11)

Post by Bramhachari Khamari » Sat Jan 09, 2021 1:27 pm

Dear Sir,
Based on your suggestions I prepare a new input file from scratch using yambo -p p -g n -V all and it created all the tags that you have highlighted

I got gw0 /ppa/ dyson/HF_and_locXC /em1d keywords at the beginning of the input file.

Then I run a serial interactive job, Calculation stops after few seconds and output the following lines

<---> [01] MPI/OPENMP structure, Files & I/O Directories
<---> [02] CORE Variables Setup
<---> [02.01] Unit cells
<---> [02.02] Symmetries
<---> [02.03] Reciprocal space
<---> [02.04] K-grid lattice
<---> Grid dimensions : 12 12
<---> [02.05] Energies & Occupations
<---> [03] Transferred momenta gridSegmentation fault (core dumped)

Regards,
Bramhachari Khamari
Bramhachari Khamari
Post Doctoral Fellow
Department of Physics
IIT Kanpur, India.

User avatar
Daniele Varsano
Posts: 4209
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: GW calculation using 5.0 fails with segmentation fault (signal 11)

Post by Daniele Varsano » Sat Jan 09, 2021 1:34 pm

Dear Bramhachari,
If you not already did, can you try to recompile the code using the --enable-int-linalg command line in configure and
see if this solves the problem?
if does not work, please post your config.log file, in order to see if there something wrong with the compilation.
if not, at this point we will need to reproduce your error to spot the problem and we need your scf/nscf/pseudo files together with the yambo input you used.
If possibile, before posting your inputs try to reproduce the error with sloppy parameters in order to make the calculation fast.

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Bramhachari Khamari
Posts: 96
Joined: Wed Oct 23, 2019 5:52 am

Re: GW calculation using 5.0 fails with segmentation fault (signal 11)

Post by Bramhachari Khamari » Sat Jan 09, 2021 1:42 pm

Dear Sir,
With new command line that you have given, I produced the required input and did a parallel calculation. Same problem appeared with segmentation fault. When I run a serial interactive job, the calculation stops after few second with the output given below

<---> [01] MPI/OPENMP structure, Files & I/O Directories
<---> [02] CORE Variables Setup
<---> [02.01] Unit cells
<---> [02.02] Symmetries
<---> [02.03] Reciprocal space
<---> [02.04] K-grid lattice
<---> Grid dimensions : 12 12
<---> [02.05] Energies & Occupations
<---> [03] Transferred momenta gridSegmentation fault (core dumped)

Regards,
Bramhachari Khamari
Bramhachari Khamari
Post Doctoral Fellow
Department of Physics
IIT Kanpur, India.

User avatar
Daniele Varsano
Posts: 4209
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: GW calculation using 5.0 fails with segmentation fault (signal 11)

Post by Daniele Varsano » Sat Jan 09, 2021 1:43 pm

Please see my post above,

Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Bramhachari Khamari
Posts: 96
Joined: Wed Oct 23, 2019 5:52 am

Re: GW calculation using 5.0 fails with segmentation fault (signal 11)

Post by Bramhachari Khamari » Sat Jan 09, 2021 1:49 pm

Dear Sir,
I attached the zip file of the config.log that is used for the compilation.

Regards,
Bramhachari Khamari
You do not have the required permissions to view the files attached to this post.
Bramhachari Khamari
Post Doctoral Fellow
Department of Physics
IIT Kanpur, India.

Bramhachari Khamari
Posts: 96
Joined: Wed Oct 23, 2019 5:52 am

Re: GW calculation using 5.0 fails with segmentation fault (signal 11)

Post by Bramhachari Khamari » Sat Jan 09, 2021 2:04 pm

Dear Sir,
I will recompile the code using --enable-int-linalg and post you further.

Regards,
Bramhachari Khamari
Bramhachari Khamari
Post Doctoral Fellow
Department of Physics
IIT Kanpur, India.

Bramhachari Khamari
Posts: 96
Joined: Wed Oct 23, 2019 5:52 am

Re: GW calculation using 5.0 fails with segmentation fault (signal 11)

Post by Bramhachari Khamari » Mon Jan 11, 2021 2:31 pm

Dear Sir,
I followed your suggestion to include the command line --enable-int-linalg and compile the code. Unfortunately I met with an error which says

ifort: error #10236: File not found: '/home/bkhamari/Software/yambo-master/lib/external/intel/mpiifort/lib/libscalapack.a'
ifort: error #10236: File not found: '/home/bkhamari/Software/yambo-master/lib/external/intel/mpiifort/lib/libblacs.a'
ifort: error #10236: File not found: '/home/bkhamari/Software/yambo-master/lib/external/intel/mpiifort/lib/libblacs_C_init.a'
ifort: error #10236: File not found: '/home/bkhamari/Software/yambo-master/lib/external/intel/mpiifort/lib/libblacs_init.a'
make[2]: *** [yambo] Error 1
make[2]: Leaving directory `/home/bkhamari/Software/yambo-master/driver'
make[1]: *** [yambo] Error 2
make[1]: Leaving directory `/home/bkhamari/Software/yambo-master'
yambo build failed

However, Without using --enable-int-linalg the code was compiled without any problem which I have mentioned earlier post. Details of the files that you asked me to sent is uploaded
as zip file.

i) Please find the config.log file with --enable-int-linalg
ii) scf/nscf/pseudo potential/yambo.in files for hBN which is used to get the segmentation fault problem during the calculation of GW. In this case I have not used ( --enable-int-linalg in config.log)
Looking forward to your further suggestions to rectify this issue.

Regards,
You do not have the required permissions to view the files attached to this post.
Bramhachari Khamari
Post Doctoral Fellow
Department of Physics
IIT Kanpur, India.

Bramhachari Khamari
Posts: 96
Joined: Wed Oct 23, 2019 5:52 am

Re: GW calculation using 5.0 fails with segmentation fault (signal 11)

Post by Bramhachari Khamari » Mon Jan 11, 2021 2:36 pm

Dear Sir,
In the previous post some of the files were not attached. Here is the rest of the files.


Regards,
You do not have the required permissions to view the files attached to this post.
Bramhachari Khamari
Post Doctoral Fellow
Department of Physics
IIT Kanpur, India.

User avatar
claudio
Posts: 528
Joined: Tue Mar 31, 2009 11:33 pm
Location: Marseille
Contact:

Re: GW calculation using 5.0 fails with segmentation fault (signal 11)

Post by claudio » Mon Jan 11, 2021 2:48 pm

Dear Bramhachari,

please have a look here, maybe it can help:

Identify what's causing segmentation fault in Yambo

best
Claudio
Claudio Attaccalite
[CNRS/ Aix-Marseille Université/ CINaM laborarory / TSN department
Campus de Luminy – Case 913
13288 MARSEILLE Cedex 09
web site: http://www.attaccalite.com

User avatar
Daniele Varsano
Posts: 4209
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: GW calculation using 5.0 fails with segmentation fault (signal 11)

Post by Daniele Varsano » Mon Jan 11, 2021 3:31 pm

Dear Bramhachari,
when adding --enable-int-linalg (internal linear algebra) you need to remove all the linking to other linear algebra libraries as
--with-blas-libs --with-lapack-libs

before recompiling be sure to clean old files:
make distclean

At this stage I would also remove the --enable-dp and parallel algebra (scalapack)

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Locked