Segmentation fault (11)

Hi

I am solving diamond example and I have installed QE 6.1 and EPW with opempi on my pc (ubuntu) and cluster (centos). In my pc this calculation goes well but in cluster it stops when I run epw.x with

[color=#000040:2pgbo1tk][i] [nadeem@CMT0 epw]$ [CMT0:16364] *** Process received signal ***
[CMT0:16364] Signal: Segmentation fault (11)
[CMT0:16364] Signal code: (128)
[CMT0:16364] Failing at address: (nil)
[CMT0:16365] *** Process received signal ***
[CMT0:16366] *** Process received signal ***
[CMT0:16366] Signal: Segmentation fault (11)
[CMT0:16366] Signal code: Address not mapped (1)
[CMT0:16366] Failing at address: 0x100000019
[CMT0:16365] Signal: Segmentation fault (11)
[CMT0:16365] Signal code: Address not mapped (1)
[CMT0:16365] Failing at address: 0x100000019
[CMT0:16363] *** Process received signal ***
[CMT0:16363] Signal: Segmentation fault (11)
[CMT0:16363] Signal code: (128)
[CMT0:16363] Failing at address: [/color:2pgbo1tk]

epw.out is reads at the end

[color=#000040:2pgbo1tk] Running Wannier90

 Wannier Function centers (cartesian, alat) and spreads (ang):

 (   0.12500   0.12500   0.12500) :   0.71730
 (   0.12500  -0.12500  -0.12500) :   0.71730
 (  -0.12500   0.12500  -0.12500) :   0.71730
 (  -0.12500  -0.12500   0.12500) :   0.71730

 -------------------------------------------------------------------
 WANNIER      :      9.77s CPU      9.80s WALL (       1 calls)
 -------------------------------------------------------------------

 Dipole matrix elements calculated


 Calculating kmap and kgmap
 Progress  kmap: ########################################
 Progress kgmap: ########################################
 kmaps        :      0.50s CPU      0.51s WALL (       1 calls)
 Symmetries of bravais lattice:  48
 Symmetries of crystal:          48


 ===================================================================
 irreducible q point #    1
 ===================================================================

 Symmetries of small group of q: 48
      in addition sym. q -> -q+G:

 Number of q in the star =    1
 List of q in the star:
      1   0.000000000   0.000000000   0.000000000
    Imposing acoustic sum rule on the dynamical matrix

    q(    1 ) = (   0.0000000   0.0000000   0.0000000 )

[/color:2pgbo1tk]

I used following execuation
in phonon folder:
mpirun -np 4 pw.x <scf.in> scf.out
mpirun -np 4 ph.x <ph.in> ph.out
python pp.py
in epw folder:
mpirun -np 4 pw.x <scf.in> scf.out
mpirun -np 4 pw.x -npool <nscf.in> nscf.out
mpirun -np 4 epw.x <epw.in> epw.out

all calculation before last step completes well. Please recommend me some solution.

Regards
Nadeem
Graduate Student
Physics, Sungkyunkwan University

Dear Nadeem,

For the epw calculation the commands should be:
mpirun -np 4 pw.x <scf.in> scf.out
mpirun -np 4 pw.x -npool 4 <nscf.in> nscf.out
mpirun -np 4 epw.x -npool 4 <epw.in> epw.out

Best,
Carla

Dear Carla
Thank your for your reply.
Actually That is typo from me. actually I did the way you mentioned in reply but got this error.

Regards
Nadeem

Dear Develops and Users
Can someone answer my query please.

Best
Nadeem

Dear Nadeem,

It seems like the code cannot access the .dvscf files produced by the phonon code. Have you run the python pp.py script correctly? It should create a "save" folder with the dvscf files and other stuff.

Best,
Samuel

Thanks for reply Sponce
Yes I ran pp.py code and it does creat save folder with following files in it
[color=#000040:1hh4zqf2]
[i:1hh4zqf2][nadeem@CMT0 save]$ ls
diam.dvscf_q1 diam.dvscf_q3 diam.dvscf_q5 diam.dvscf_q7 diam.dyn_q1 diam.dyn_q3 diam.dyn_q5 diam.dyn_q7 diam.phsave
diam.dvscf_q2 diam.dvscf_q4 diam.dvscf_q6 diam.dvscf_q8 diam.dyn_q2 diam.dyn_q4 diam.dyn_q6 diam.dyn_q8
[nadeem@CMT0 save]$
[/i:1hh4zqf2][color=#000040][/color:1hh4zqf2]

in
../save/diam.phsave
[i:1hh4zqf2][nadeem@CMT0 diam.phsave]$ ls
control_ph.xml dynmat.2.1.xml dynmat.3.1.xml dynmat.4.1.xml dynmat.5.1.xml dynmat.5.6.xml dynmat.6.4.xml dynmat.7.2.xml dynmat.8.3.xml patterns.5.xml
dynmat.1.0.xml dynmat.2.2.xml dynmat.3.2.xml dynmat.4.2.xml dynmat.5.2.xml dynmat.6.0.xml dynmat.6.5.xml dynmat.7.3.xml patterns.1.xml patterns.6.xml
dynmat.1.1.xml dynmat.2.3.xml dynmat.3.3.xml dynmat.4.3.xml dynmat.5.3.xml dynmat.6.1.xml dynmat.6.6.xml dynmat.8.0.xml patterns.2.xml patterns.7.xml
dynmat.1.2.xml dynmat.2.4.xml dynmat.3.4.xml dynmat.4.4.xml dynmat.5.4.xml dynmat.6.2.xml dynmat.7.0.xml dynmat.8.1.xml patterns.3.xml patterns.8.xml
dynmat.2.0.xml dynmat.3.0.xml dynmat.4.0.xml dynmat.5.0.xml dynmat.5.5.xml dynmat.6.3.xml dynmat.7.1.xml dynmat.8.2.xml patterns.4.xml status_run.xml
[nadeem@CMT0 diam.phsave]$
[/i:1hh4zqf2]

Regards

Hello,

Which compiler are you using?. Could you recompile the code with debugging flags.
Can you then try to run it with gdb to see at which line of the code the segfault occurs. At the moment the segfault message is not very helpful.

Best,
Samuel

Thank you Samuel

My mechine has "gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-18)". And I am installing EPW-4.1.0. Is this compiler incompatible? Sorry I tried to learn how to run my program with debugger but I could not.

Regards
Nadeem

Dear Nadeem,

It should be ok. You can try with a slightly more recent version of GCC.

The most important is to have "-g" flags when you compile (in the make.inc file).

With this you can then try to run it with gdb to see where the segfault happen. If you tell me where I should be able to help.

Best,
Samuel

Hi Samuel

Now I compiled EPW-4.1 with gcc-5.2.0 and I ran all the tests in ../qe-6.0/EPW/tests/

All the test calculations ran well exept epw2.in and epw3.in in test01. Output says

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Error in routine epw_init (1):
Error: kmaps has to be true for a restart run.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

would you please explain this error? Thank you for help

Regards
Nadeem

Dear Nadeem,

Happy to see that it now works for you.

The EPW/tests are not the official tests (its a remain of previous tests and will eventually be removed). In this case you just need to set "kmaps = .true." in epw2.in and epw3.in to make it works (as the error message tells us).

Note that the version 6.1 of quantum espresso is available: http://www.qe-forge.org/gf/project/q-e/ … kage_id=18

Best,
Samuel

Hi Samuel
Thank You for so much Help !!

Dear users,
I am using epw.x to examine electron phonon coupling in PbI2. already I have gone through tutorial for SiC. But for PbI2 ,as soon as the wannier part gets completed, I get an error
"segmentation fault occurred " . I am giving my epw input file here. can you please help me to fix it ? thanking you in advance

–
&inputepw
prefix =‘pbi2’,
amass(1) = 207.20000
amass(2) = 126.90447
outdir = ‘./’
dvscf_dir = ‘./save’

elph = .true.
kmaps = .false.
epbwrite = .true.
epbread = .false.
epwwrite = .true.
epwread = .false.
lpolar = .true.

wannierize = .true.
system_2d = .true.
nbndsub = 12
nbndskip = 0

efermi_read= .true.
fermi_energy= -1.03385

dis_win_max = 10.0d0
dis_froz_max = 0.0d0
dis_win_min = -5.0d0
iprint = 2

proj(1)= ‘Pb:s’
proj(2)= ‘Pb:p’
proj(3)= ‘I:s’
proj(4)= ‘I:p’
wdata(1) = ‘bands_plot = .true.’
wdata(2) = ‘begin kpoint_path’
wdata(3) = ’ K 0.3333333333 0.3333333333 0.00000 G 0.00000 0.00000 0.00000 ’
wdata(4) = ’ G 0.00000 0.00000 0.00000 M 0.50000 0.00000 0.00000 ’
wdata(5) = ’ M 0.50000 0.00000 0.00000 K 0.3333333333 0.3333333333 0.00000 ’
wdata(6) = ‘end kpoint_path’

elecselfen = .false.
phonselfen = .false.
prtgkk = .true.

a2f = .false.
parallel_k = .true.
parallel_q = .false.
fsthick = 7 ! eV
eptemp = 300 ! K
degaussw = 0.05 ! eV

filqf = ‘path.dat’
nkf1 = 1
nkf2 = 1
nkf3 = 1
!nqf1 = 90
!nqf2 = 90
!nqf3 = 1
nk1 = 15
nk2 = 15
nk3 = 1
nq1 = 5
nq2 = 5
nq3 = 1
/
5
0.000000000000000 0.000000000000000 0.000000000000000
0.000000000000000 0.230972210747074 0.000000000000000
0.000000000000000 0.461944421494147 0.000000000000000
0.200027802064237 0.346458316236113 0.000000000000000
0.200027802064237 0.577430526983186 0.000000000000000

I have somethng more to ask.
In my nscf file , I have given nbnd=35, but I am giving only 12 wannier band to plot the band structure and putting nbndskip=0,
is error because of this ? I dont know , please help me

Hello,

Yes it could be that there is a problem at the Wannierization level.

You can check the output of Wannier. The file should be named pbi2.wout. Did it converged ?

Best wishes,
Samuel

thanks for your reply
Actually, one of individual spread was high around 14 (ang. square ) . I am doing it with changing some parameters. I will let you know, once it is done.

Dear users,
I am getting a weird thing in my epw calculation.
When I give uniform fine grid in epw input file to calculate |g|, it is smoothly running, but when I am giving the desired path,
it is giving me an error related to the segmentation fault.
one more thing, after it runs perfectly where I use only uniform fine grid, If I restart this calculation to calculate Imsigma (even if I use unform fine grid) it gives me again the same error.
Please Help me out . thanking you in advance

I am giving you my input file

–
&inputepw
prefix = ‘pbi2’ ,
amass(1) = 207.20000
amass(2) = 126.90447
outdir = ‘./’
dvscf_dir = ‘./save’
elph = .true.
kmaps = .false.
epbwrite = .true.
epbread = .false.
epwwrite = .true.
epwread = .false.
lpolar = .true.
wannierize = .true.
system_2d = .true.
nbndsub = 12
nbndskip = 0

efermi_read= .true.
fermi_energy= -1.03385

dis_win_max = 10.0d0
dis_froz_max = 0.0d0
dis_win_min = -5.0d0
iprint = 2

num_iter = 5000

proj(1)= ‘Pb:s’
proj(2)= ‘Pb:p’
proj(3)= ‘I:s’
proj(4)= ‘I:p’

elecselfen = .false.
phonselfen = .false.
prtgkk = .true.

a2f = .false.
parallel_k = .true.
parallel_q = .false.
fsthick = 8 ! eV
eptemp = 300 ! K
degaussw = 0.05 ! eV

filqf =‘path.dat’
nkf1 = 1
nkf2 = 1
nkf3 = 1
nk1 = 15
nk2 = 15
nk3 = 1
nq1 = 5
nq2 = 5
nq3 = 1
/
5 cartesian
0.000000000000000 0.000000000000000 0.000000000000000
0.000000000000000 0.230972210747074 0.000000000000000
0.000000000000000 0.461944421494147 0.000000000000000
0.200027802064237 0.346458316236113 0.000000000000000
0.200027802064237 0.577430526983186 0.000000000000000

and the error is

forrtl: severe (151): allocatable array is already allocated
Image PC Routine Line Source
epw.x 0000000000E957DC Unknown Unknown Unknown
epw.x 0000000000445E15 ephwann_shuffle_ 234 ephwann_shuffle.f90
epw.x 0000000000442207 elphon_shuffle_wr 764 elphon_shuffle_wrap.f90
epw.x 000000000040A185 MAIN__ 150 epw.f90
epw.x 00000000004094DE Unknown Unknown Unknown
libc-2.12.so 00002ACFAE299D1D libc_start_main Unknown Unknown
epw.x 00000000004093E9 Unknown Unknown Unknown
forrtl: error (69): process interrupted (SIGINT)
Image PC Routine Line Source
epw.x 0000000000E4EB30 Unknown Unknown Unknown
libpthread-2.12.s 00002B0B289BB710 Unknown Unknown Unknown
libmpi.so.12 00002B0B27EE127D PMPIDI_CH3I_Progr Unknown Unknown
libmpi.so.12 00002B0B27FEDC29 MPIC_Wait Unknown Unknown
libmpi.so.12 00002B0B27FED922 MPIC_Sendrecv Unknown Unknown
libmpi.so.12 00002B0B27EC6E55 MPIR_Barrier_intr Unknown Unknown
libmpi.so.12 00002B0B27EC6875 I_MPIR_Barrier_in Unknown Unknown
libmpi.so.12 00002B0B27EC66CC MPIR_Barrier Unknown Unknown
libmpi.so.12 00002B0B27EC65FC MPIR_Barrier_impl Unknown Unknown
libmpi.so.12 00002B0B27EC8482 PMPI_Barrier Unknown Unknown
libmpifort.so.12. 00002B0B27A8F38C pmpi_barrier Unknown Unknown
epw.x 0000000000C2F6E7 bcast_integer
37 mp_base.f90
epw.x 0000000000C13888 mp_mp_mp_bcast_i1 293 mp.f90
epw.x 000000000044602E ephwann_shuffle
243 ephwann_shuffle.f90
epw.x 0000000000442207 elphon_shuffle_wr 764 elphon_shuffle_wrap.f90
epw.x 000000000040A185 MAIN__ 150 epw.f90
epw.x 00000000004094DE Unknown Unknown Unknown
libc-2.12.so 00002B0B28E6BD1D __libc_start_main Unknown Unknown
epw.x 00000000004093E9 Unknown Unknown Unknown

Dear vinpandey,

Seems there is an issue with the EPW version you are using (array already allocated). Which version are you using and what is at (and around) line 234 of the q-e/EPW/src/ephwann_shuffle.f90 file ?
In the most recent version of EPW it is just a variable declaration so its probably not that.

PS: Can you try with the most recent QE/EPW version that you can get with:
git clone https://gitlab.com/QEF/q-e.git

Best wishes,
Samuel