Segmentation fault during DMFT/QMC

Total energy, geometry optimization, DFT+U, spin....

Moderator: bguster

Locked
PhilippBa
Posts: 15
Joined: Tue May 26, 2015 7:22 pm

Segmentation fault during DMFT/QMC

Post by PhilippBa » Thu Aug 06, 2015 6:33 pm

Dear all,


I am using the CT-QMC solver from the ABINIT DMFT implementation to investigate a single Cobalt atom on a copper surface.


So far, the code runs fine for temperatures about ß=100ev^-1 or ß=130eV^-1, as long as I do not use too much time slices for the impurity Green function. At the moment i use 200 and this should be fine for this temperature, but if i decrease increase ß to 200 eV^-1 I have to increase the slices, too.

But in this case the Code runs into some segmentation fault.

For testing purposes i decresed the original system size (over 90 atoms to 55 atoms) but used high values for the time slices instead (1000). So i can reproduce the segmentation fault much faster.

Now more details to the problem: It seems to occur during the calculation of the log frequencies:

====================================
===== Start of DMFT calculation
====================================

=> Seed for QMC inside DMFT is dmftqmc_seed= 0
...

--ifreq-- 1554 0.13063327120096E+02 0.38858160054057E+02
--ifreq-- 1555 0.13153379220438E+02 0.39556089269338E+02
--ifreq-- 1556 0.13245740348994E+02 0.39933746960086E+02
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
libirc.so 00007F1926FA5961 Unknown Unknown Unknown
libirc.so 00007F1926FA40B7 Unknown Unknown Unknown
libifcore.so.5 00007F19289547B2 Unknown Unknown Unknown
libifcore.so.5 00007F1928954606 Unknown Unknown Unknown
libifcore.so.5 00007F19288C142C Unknown Unknown Unknown
libifcore.so.5 00007F19288D2328 Unknown Unknown Unknown
libpthread.so.0 00007F1928DF2130 Unknown Unknown Unknown
abinit 00000000021AA505 Unknown Unknown Unknown
abinit 0000000000D39A85 Unknown Unknown Unknown
abinit 00000000005FB00B Unknown Unknown Unknown
abinit 0000000000599F54 Unknown Unknown Unknown
abinit 000000000058FE18 Unknown Unknown Unknown
abinit 0000000000533DCD Unknown Unknown Unknown
abinit 000000000042EC81 Unknown Unknown Unknown
abinit 00000000004137F5 Unknown Unknown Unknown
abinit 000000000040875F Unknown Unknown Unknown
abinit 00000000004067CE Unknown Unknown Unknown
libc.so.6 00007F1926BF9AF5 Unknown Unknown Unknown
abinit 00000000004066D9 Unknown Unknown Unknown

...



As far as i know, such faults can be due to low memory allocation. However, I am not an expert on this. I hope that one of you can give me instructions how to debug this problem. So far, I could not find an error in the code (even different compiling settings did not helped).

I am happy with any suggestions!


Best regards,


Philipp



PS: I attached the input file of my calulation. The LDA input is almost the same but without the DMFT and ldA+U settings (and nstep = 120)
Attachments
co.in
(2.8 KiB) Downloaded 276 times
slurm-725.out
(342.95 KiB) Downloaded 275 times
Philipp Bahlke
University of Hamburg
Institute for inorganic and applied chemistry
Germany

Locked