[QE-developers] Problem related to dos.x and bands.x

[Sujan Budhathoki]

Hi developers,

I am using QE6.1.
To be precise, I can run SCF, NSCF and the relaxation runs pretty smoothly
without any problem.
However, each time:
(1) If use dos.x to get density of states after the NSCF run, I get an
error:

[compute-5-15:41909] Signal: Segmentation fault (11)
[compute-5-15:41909] Signal code: Address not mapped (1)
[compute-5-15:41909] Failing at address: (nil)
[compute-5-15:41909] [ 0] /lib64/libpthread.so.0(+0xf5e0)[0x2ac67aeeb5e0]
[compute-5-15:41909] [ 1] /lib64/libc.so.6(+0x152c20)[0x2ac67b24ac20]
[compute-5-15:41909] [ 2]
/share/apps/openmpi/openmpi1.10.2gcc/lib/libopen-pal.so.13(opal_convertor_pack+0x196)[0x2ac67bb8cf46]
[compute-5-15:41909] [ 3]
/share/apps/openmpi/openmpi1.10.2gcc/lib/openmpi/mca_btl_sm.so(mca_btl_sm_sendi+0x2c5)[0x2ac6819e9515]
[compute-5-15:41909] [ 4]
/share/apps/openmpi/openmpi1.10.2gcc/lib/openmpi/mca_pml_ob1.so(+0x805e)[0x2ac68201905e]
[compute-5-15:41909] [ 5]
/share/apps/openmpi/openmpi1.10.2gcc/lib/openmpi/mca_pml_ob1.so(mca_pml_ob1_isend+0xaf)[0x2ac68201933f]
[compute-5-15:41909] [ 6]
/share/apps/openmpi/openmpi1.10.2gcc/lib/openmpi/mca_coll_tuned.so(ompi_coll_tuned_bcast_intra_generic+0x352)[0x2ac682e8b342]
[compute-5-15:41909] [ 7]
/share/apps/openmpi/openmpi1.10.2gcc/lib/openmpi/mca_coll_tuned.so(ompi_coll_tuned_bcast_intra_binomial+0xd8)[0x2ac682e8b758]
[compute-5-15:41909] [ 8]
/share/apps/openmpi/openmpi1.10.2gcc/lib/openmpi/mca_coll_tuned.so(ompi_coll_tuned_bcast_intra_dec_fixed+0xcc)[0x2ac682e80abc]
[compute-5-15:41909] [ 9]
/share/apps/openmpi/openmpi1.10.2gcc/lib/libmpi.so.12(MPI_Bcast+0x130)[0x2ac67a420310]
[compute-5-15:41909] [10]
/share/apps/openmpi/openmpi1.10.2gcc/lib/libmpi_mpifh.so.12(pmpi_bcast+0x7e)[0x2ac67a1b5e9e]
[compute-5-15:41909] [11] /share/apps/espresso/bin/dos.x[0x5e9ed2]
[compute-5-15:41909] [12] /share/apps/espresso/bin/dos.x[0x5db515]
[compute-5-15:41909] [13] /share/apps/espresso/bin/dos.x[0x4d5f0e]
[compute-5-15:41909] [14] /share/apps/espresso/bin/dos.x[0x4e0508]
[compute-5-15:41909] [15] /share/apps/espresso/bin/dos.x[0x407593]
[compute-5-15:41909] [16] /share/apps/espresso/bin/dos.x[0x404679]
[compute-5-15:41909] [17] /share/apps/espresso/bin/dos.x[0x7789ea]
[compute-5-15:41909] [18]
/lib64/libc.so.6(__libc_start_main+0xf5)[0x2ac67b119c05]
[compute-5-15:41909] [19] /share/apps/espresso/bin/dos.x[0x4044f9]
[compute-5-15:41909] *** End of error message ***
srun: error: compute-5-15: task 0: Segmentation fault
srun: First task exited 60s ago
srun: step:1569625.0 tasks 1-7: running
srun: step:1569625.0 task 0: exited abnormally
srun: Terminating job step 1569625.0
srun: Job step aborted: Waiting up to 62 seconds for job step to finish.
slurmstepd: error: *** STEP 1569625.0 ON compute-5-15 CANCELLED AT
2020-08-18T00:47:37 ***
srun: error: compute-20-1: task 7: Killed
srun: error: compute-5-15: tasks 1-4: Killed
srun: error: compute-20-0: tasks 5-6: Killed

(3) If I run dos.x directly after SCF run then I get no error.
(2) I get the same error if I use bands.x  [No problem to run: pw.x <
Si_bands.in > Si_bands.out]

 I inspected the usage of memory and it looks fine. The input data file is
okay.

Any suggestions will be highly appreciated.


sincerely,
sujan budhathoki
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20200820/8fc845bb/attachment.html>