em1s and bse -> Segmentation Fault

[Stephan]

Hi everybody,

I try to calculate optical properties with the bse-runlevel.
I follow the instructions in the "Solid LiF:excitons at work" -tutorial on the homepage.
At first I calculate bthe statically screened electron-electron interaction with the input-file:

em1s # [R Xs] Static Inverse Dielectric Matrix
FFTGvecs= 11701 RL # [FFT] Plane-waves
Chimod= "hartree" # [X] IP/Hartree/ALDA/LRC/BSfxc
% QpntsRXs
1 | 30 | # [Xs] Transferred momenta
%
% BndsRnXs
11 | 36 | # [Xs] Polarization function bands
NGsBlkXs= 311 RL # [Xs] Response block size
% LongDrXs
0.1000E-4 | 0.000 | 0.000 | # [Xs] [cc] Electric Field
%

The l_ and r_ file you will find in the attachment. The programm does not complain
When I continue now with yambo -o b -k sex -y h

I receive the input file:
# ____ ____ _ ____ ____ ______ ___
# |_ _||_ _| / \ |_ \ / _||_ _ \ ." `.
# \ \ / / / _ \ | \/ | | |_) | / .-. \
# \ \/ / / ___ \ | |\ /| | | __". | | | |
# _| |_ _/ / \ \_ _| |_\/_| |_ _| |__) |\ `-" /
# |______||____| |____||_____||_____||_______/ `.___."
#
# GPL Version 3.4.1 Revision 3187
# http://www.yambo-code.org
#
optics # [R OPT] Optics
bse # [R BSE] Bethe Salpeter Equation.
bsk # [R BSK] Bethe Salpeter Equation kernel
bss # [R BSS] Bethe Salpeter Equation solver
em1s # [R Xs] Static Inverse Dielectric Matrix
FFTGvecs= 11701 RL # [FFT] Plane-waves
Chimod= "Hartree" # [X] IP/Hartree/ALDA/LRC/BSfxc
BSEmod= "causal" # [BSE] resonant/causal/coupling
BSKmod= "SEX" # [BSE] IP/Hartree/HF/ALDA/SEX
BSSmod= "h" # [BSS] (h)aydock/(d)iagonalization/(i)nversion/(t)ddft`
BSENGexx= 65617 RL # [BSK] Exchange components
BSENGBlk= 311 RL # [BSK] Screened interaction block size
#WehCpl # [BSK] eh interaction included also in coupling
% BEnRange
0.00000 | 10.00000 | eV # [BSS] Energy range
%
% BDmRange
0.10000 | 0.10000 | eV # [BSS] Damping range
%
BEnSteps= 100 # [BSS] Energy steps
% BLongDir
1.000000 | 0.000000 | 0.000000 | # [BSS] [cc] Electric Field
%
% BSEBands
1 | 45 | # [BSK] Bands range
%
BSHayTrs= -0.02000 # [BSS] [o/o] Haydock treshold. Strict(>0)/Average(<0)
% QpntsRXs
1 | 30 | # [Xs] Transferred momenta
%
% BndsRnXs
11 | 36 | # [Xs] Polarization function bands
%
NGsBlkXs= 311 RL # [Xs] Response block size
% LongDrXs
0.1000E-4 | 0.000 | 0.000 | # [Xs] [cc] Electric Field
%

When I try to run the code the calculation cancels with the error message:

[QuanQal2:mpi_rank_1][error_sighandler] Caught error: Segmentation fault (signal 11)
[QuanQal2:mpi_rank_2][error_sighandler] Caught error: Segmentation fault (signal 11)
[QuanQal2:mpi_rank_3][error_sighandler] Caught error: Segmentation fault (signal 11)
[QuanQal2:mpi_rank_0][error_sighandler] Caught error: Segmentation fault (signal 11)

===================================================================================
= BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
= PID 26291 RUNNING AT QuanQal2
= EXIT CODE: 139
= CLEANING UP REMAINING PROCESSES
= YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================
YOUR APPLICATION TERMINATED WITH THE EXIT STRING: Segmentation fault (signal 11)
This typically refers to a problem with your application.
Please see the FAQ page for debugging suggestions

The l_ file of this run is also in the attachment. The r_file will follow.
I already tried to reduce the computational cost by reducing the considered bands FFTvecs etc. The Segmentation fault remains.
In the tutorial it is said that after the em1s-run I should have a SAVE/ndb.em1s database. Thats not the case I just have the file SAVE/db.ems.

Can anybody please help me and tell me what's wrong?

Thanks and Regards

Stephan

[Daniele Varsano]

Dear Stephan,

In the tutorial it is said that after the em1s-run I should have a SAVE/ndb.em1s database. Thats not the case I just have the file SAVE/db.ems.

This is because in the compilation of the coded you did not linked the netcdf libraries.
Although it is mandatory for the 4.0 release, it is not for the 3.4 you are using. So this is not the problem.

Even if not totally sure as the eerror you posted is produced by the scheduler and not the code.
Anyway, after a quick look to your input/reports I can see:

[RD./SAVE//db.kindx]----------------------------------------
Polarization last K : 244
QP states : 1 244
X grid is uniform :no
BS scattering :yes
- S/N 008162 --------------------------- v.03.04.01 r.3187 -

IBZ Q-points : 1
BZ Q-points : 1

You have many K points, but only 1 q point (gamma) has been considered as your X grid is not uniform.
This is a problem because your screening has been calculated for a q point only and the BSE need all the q point in the Bz.

So my suggestion is to understand first why your K point sampling is not uniform.

Next, some suggestions:

For the calculation of the screening it i a good idea to use all the occupied bands:

Code: Select all

% BndsRnXs
1 | Nbands | # [Xs] Polarization function bands
%

On the other hand this is not the case for the BSE, where you can have converged results considering a windows of bands around the fermi energy.
Now looking at your input, the dimension of your BSE Hamiltonian: (Nc x Nv x Nk is bigger than 1 million), which probably is beyond your computational power.

Next:

Code: Select all

% QpntsRXs
1 | 30 | # [Xs] Transferred momenta
%

It is not a good idea to limit the calculation of the screening to a small number of Q as it is needed for all the Q in the BZ.
Finally please note you have a metallic system is that correct?

In any case the first think you need to understand is why your k grid it is not uniform. In the ground state did you use an automatic grid or you supplied it? If you are using quantum espresso, and an automatic grid, the use of force_symmorphic=.true. flag can help to solve the issue.

Best,
Daniele