Dipole moment segmentation fault

[Vladislav Sláma]

Dear all,

I’m trying to simulate optical spectra and plasmon resonance for different gold cluster sizes. In order to do that, I’m employing real time propagation after initial delta-pulse kick. For the smaller particle sizes (up to ~114 atoms) everything works well and I get reasonable spectra. However for the larger nanoparticle (178 atoms) I’m getting segmentation fault due to the calculation of the dipole moments. When I do not compute and print dipole moments, the real time propagation goes well without any problem. Could you please suggest me how to compute the dipole moment during the real time propagation to avoid this problem?

Thank you. Best wishes,

Vladislav

[Krack Matthias]

Hi Vladislav

 

Segmentation faults are often caused by an insufficient stack size. Try to set (assuming bash) “ulimit -s unlimited” and “export OMP_STACKSIZE=16M” before launching the cp2k run.

 

HTH

 

Matthias

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/f1ad08c5-e388-4e1a-bbbc-fbd46e715bf6n%40googlegroups.com.

[Vladislav Sláma]

Dear Matthias,

Thank you for your suggestion. I was already using “ulimit -s unlimited”, but adding “export OMP_STACKSIZE=16M” didn’t help and I’m still getting the same segmentation fault. I’m including my submission script on CSCS cluster to provide all the settings for the cp2k run.

Best wishes,

Vladislav

Dne středa 24. května 2023 v 9:06:47 UTC+2 uživatel Krack Matthias napsal:

[Vladislav Sláma]

Hi, 

I'm including a small update from the run on different machine with differently compiled version of CP2K which provides a more detailed information about the source of the error:

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0  0x628b4ef in ???
        at /home/abuild/rpmbuild/BUILD/glibc-2.31/nptl/../sysdeps/unix/sysv/linux/x86_64/sigaction.c:0
#1  0xc9b71e in __qs_moments_MOD_build_local_moment_matrix
        at /home/krack/github/cp2k/cp2k-2023.1/src/qs_moments.F:253
#2  0xca0074 in __qs_moments_MOD_qs_moment_locop
        at /home/krack/github/cp2k/cp2k-2023.1/src/qs_moments.F:1874
#3  0xb923fe in __qs_scf_post_gpw_MOD_qs_scf_post_moments
        at /home/krack/github/cp2k/cp2k-2023.1/src/qs_scf_post_gpw.F:1262
#4  0xb93d18 in __qs_scf_post_gpw_MOD_scf_post_calculation_gpw
        at /home/krack/github/cp2k/cp2k-2023.1/src/qs_scf_post_gpw.F:376
#5  0x18ba70d in __qs_scf_post_scf_MOD_qs_scf_compute_properties
        at /home/krack/github/cp2k/cp2k-2023.1/src/qs_scf_post_scf.F:78
#6  0x1823e51 in __qs_scf_MOD_scf
        at /home/krack/github/cp2k/cp2k-2023.1/src/qs_scf.F:301
#7  0x11adb1b in __qs_energy_MOD_qs_energies
        at /home/krack/github/cp2k/cp2k-2023.1/src/qs_energy.F:111
#8  0xe2f6c0 in __force_env_methods_MOD_force_env_calc_energy_force
        at /home/krack/github/cp2k/cp2k-2023.1/src/force_env_methods.F:259
#9  0x5282c7 in rt_initial_guess
        at /home/krack/github/cp2k/cp2k-2023.1/src/motion/rt_propagation.F:460
#10  0x5282c7 in __rt_propagation_MOD_rt_prop_setup
        at /home/krack/github/cp2k/cp2k-2023.1/src/motion/rt_propagation.F:131
#11  0x478b8f in cp2k_run
        at /home/krack/github/cp2k/cp2k-2023.1/src/start/cp2k_runs.F:383
#12  0x47b672 in __cp2k_runs_MOD_run_input
        at /home/krack/github/cp2k/cp2k-2023.1/src/start/cp2k_runs.F:997
#13  0x468771 in cp2k
        at /home/krack/github/cp2k/cp2k-2023.1/src/start/cp2k.F:379
#14  0x405a4c in main
        at /home/krack/github/cp2k/cp2k-2023.1/src/start/cp2k.F:44
/var/spool/slurm/job299631/slurm_script: line 17:  1360 Segmentation fault      /home/vslama/software/cp2k-2023.1-Linux-gnu-x86_64.ssmp -i Au_NP_small_OT_RTP_z.in -o Au_NP_small_OT_RTP_z.out

I'll be glad for any suggestion how to overcome this error.

Best wishes,

Vladislav

Dne středa 24. května 2023 v 9:29:57 UTC+2 uživatel Vladislav Sláma napsal:

[Krack Matthias]

I suggest to open an issue on github for this and to address the author of the routine.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/9b03e074-7bc9-4b4a-be26-54b0dc37f1e8n%40googlegroups.com.