Segmentation fault occurred in sample input (/v3/t78.in)

Hiroki · Post by **Hiroki** » Mon Nov 19, 2018 2:02 am

Dear developers and users,

I'm trying to calculate the optical conductivity for metals with conducti modules.
According to the conducti manual, the 3 steps are required to calculate the conductivity:
1. SCF calculation (iscf=3)
2. non-SCF calculation (iscf=-2)
3. response function calculations (iscf=-3)

First of all, I tried to calculate with the sample input.
When I had executed the sample input file for the calculation of response functions (v3/t78.in), the segmentation fault occurred in the RF step(>DS3).
The last lines in the log file are following:

Code: Select all

At SCF step   12   max residual=  9.42E-23 < tolwfr=  1.00E-22 =>converged.
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image              PC                Routine            Line        Source
libifcoremt.so.5   00002AAAAC642F0C  for__signal_handl     Unknown  Unknown
libpthread-2.17.s  00002AAAAEB40680  Unknown               Unknown  Unknown
libblas.so.3.4.2   00002AAAAB464AAC  zdotc_                Unknown  Unknown
abinit             000000000212CBF7  Unknown               Unknown  Unknown
abinit             0000000000925747  Unknown               Unknown  Unknown
abinit             00000000007A26BF  Unknown               Unknown  Unknown
abinit             0000000000520824  Unknown               Unknown  Unknown
abinit             0000000000457028  Unknown               Unknown  Unknown
abinit             000000000041576C  Unknown               Unknown  Unknown
abinit             000000000040A258  Unknown               Unknown  Unknown
abinit             00000000004079CE  Unknown               Unknown  Unknown
libc-2.17.so       00002AAAAF32C3D5  __libc_start_main     Unknown  Unknown
abinit             00000000004078C9  Unknown               Unknown  Unknown

SCF(DS1) and non-SCF(DS2) steps were terminated and output files were created successfully.

Why does this problem happen? I was trying to solve this problem and looking for the post regarding this matter, but I couldn't solve it...
ABINIT version is 8.8.4, and the executable is parallelized.
For parallel version, I used "autoparal 1" in the sample file.
The used pseudopotential is "Al.GGA_PBE-JTH.xml", which is distributed in ABINIT web.
Here is the compile option:

Code: Select all

=== Build Information === 
  Version       : 8.8.4
  Build target  : x86_64_linux_intel18.0
  Build date    : 20181110
 
 === Compiler Suite === 
  C compiler       : intel18.0
  C++ compiler     : gnu4.9
  Fortran compiler : intel18.0
  CFLAGS           :  -g -O2 -vec-report0
  CXXFLAGS         :  -g -O2 -mtune=native -march=native
  FCFLAGS          :  -g -extend-source -noaltparam -nofpscomp  -qopenmp
  FC_LDFLAGS       :    -static-intel -static-libgcc
 
 === Optimizations === 
  Debug level        : basic
  Optimization level : standard
  Architecture       : intel_xeon
 
 === Multicore === 
  Parallel build : yes
  Parallel I/O   : auto
  openMP support : yes
  GPU support    : no
 
 === Connectors / Fallbacks === 
  Connectors on : yes
  Fallbacks on  : yes
  DFT flavor    : none
  FFT flavor    : none
  LINALG flavor : netlib
  MATH flavor   : none
  TIMER flavor  : abinit
  TRIO flavor   : none
 
 === Experimental features === 
  Bindings            : @enable_bindings@
  Exports             : no
  GW double-precision : no

Sincerely yours,

Post by **jbeuken** » Wed Nov 21, 2018 8:45 am

can you try without mpi and with mkl as linalg ?

jmb

Hiroki · Post by **Hiroki** » Thu Nov 22, 2018 10:21 am

Dear Jbeuken,
Thank you for your reply.

I tried to compile using 'with_linalg_flavor=mkl', I got the following:

Code: Select all

 * LINALG flavor = netlib-fallback (libs: ignored)

I couldn't compile abinit with mkl...

Anyway, I'd like to compile the parallel version.
Which option do I choose together with mpi and openmp to calculate the response function without SIGSEG?
Or, mkl has to be required?

Sincerely yours
Hiroki,

Hiroki · Post by **Hiroki** » Wed Nov 28, 2018 11:43 pm

Dear Jbeuken,

I could compile ABINIT with mkl as linalg(also with mpi).
And I tried to calculate the sample input, then the calculation have done successfully.

Thank you very much for your help!
Hiroki,

ABINIT Discussion Forums

Segmentation fault occurred in sample input (/v3/t78.in) [SOLVED]

Segmentation fault occurred in sample input (/v3/t78.in)

Re: Segmentation fault occurred in sample input (/v3/t78.in)

Re: Segmentation fault occurred in sample input (/v3/t78.in)

Re: Segmentation fault occurred in sample input (/v3/t78.in) [SOLVED]