[QE-users] Segmentation fault with ph.x in QE 6.6.0

[pboulet]

Dear all,

I am facing a strange problem with ph.x. I am running QE 6.6.0.

I have a hexagonal structure with 9 atoms. Now:
- When I calculate phonons at any q-points but Gamma, the job runs fine.
- When I calculate phonons at Gamma without epsil (=.true.), the job runs fine.
- When I calculate phonons at Gamma with epsil=.true., I encounter a segmentation fault.

I have run the test for AlAs, the calculation has run fine.

I have read in a post that it could be due to the size of the problem, but I doubt this is the case here.

Does anyone have ever encountered such a behavior?

Here are some technical details:
QE is compiled with intel/20.0.0, mkl/20.0.0, openMPI/4.0, elpa/2018.11.001, fox/4.1.2, hdf5/1.12.0

I am running the calculations with USPP and PBE, Ecut=50Ry, and EcutRho=500Ry.
I have run a very tight optimization: Econv=1e-9, forc_conv=1e-6, conv_thr=1e-12, press_conv_thr=0.01

The job stops  just after the printing of the dielectric matrix in the output file:
>>>
iter #  15 total cpu time :  1379.5 secs   av.it.:  20.2
      thresh= 1.085E-08 alpha_mix =  0.200 |ddv_scf|^2 =  6.133E-15

     End of electric fields calculation

End of electric fields calculation

          Dielectric constant in cartesian axis

          (      99.611956135       0.000000001       0.000000000 )
          (       0.000000001      99.611956151       0.000000000 )
          (      -0.000000000      -0.000000000      46.907611459 )
<<<


The xx/yy values seem very high to me… They should be around 55. I do not know the reason. The zz value looks ok. Any idea?

Here is a small sample of the log file with the error:
Caught signal 11 (Segmentation fault: Sent by the kernel at address (nil))
==== backtrace (tid:  46520) ====
 0 0x0000000000053755 ucs_debug_print_backtrace()  /tmp/ucx/1.9.0/system-system/ucx-1.9.0/src/ucs/debug/debug.c:656
 1 0x000000000012b231 rml::internal::LocalLOCImpl<8, 32>::get()  /nfs/site/proj/openmp/promo/20190814/cet-enabled/tbbmalloc/build/linux_intel64_icc_cc4.4.6_libc2.12_kerne
l2.6.32_release/../../src/tbbmalloc/frontend.cpp:2229
 2 0x000000000012b231 rml::internal::MemoryPool::getFromLLOCache()  /nfs/site/proj/openmp/promo/20190814/cet-enabled/tbbmalloc/build/linux_intel64_icc_cc4.4.6_libc2.12_ke
rnel2.6.32_release/../../src/tbbmalloc/frontend.cpp:2271


Thank you for your help and time,

Pascal


Pascal Boulet
—
Professor in computational materials chemistry - DEPARTMENT OF CHEMISTRY
University of Aix-Marseille - Avenue Escadrille Normandie Niemen - F-13013 Marseille - FRANCE
Tél: +33(0)4 13 55 18 10 - Fax : +33(0)4 13 55 18 50
Email : pascal.boulet at univ-amu.fr <mailto:pascal.boulet at univ-amu.fr>




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20220519/8690d4e4/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 1269 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20220519/8690d4e4/attachment.p7s>