Page 1 of 1
Vasp 6.2.1 Intel Oneapi Compiler 2021.3.0 /Openmpi 4.1.1 nonl.f90: catastrophic error
Posted: Mon Aug 09, 2021 7:01 am
by hayak
Hi,
I try to compile vasp 6.2.1 using Intel OneApi Compiler 2021.3.0 + OpenMPI 4.1.1 (makefile.include.linux_intel_omp) and get with error:
nonl.f90: catastrophic error: **Internal compiler error: segmentation violation signal raised** Please report this error along with the circumstances in which it occurred in a Software Problem Report. Note: File and line given may not be explicit cause of this error.
compilation aborted for nonl.f90 (code 1)
Re: Vasp 6.2.1 Intel Oneapi Compiler 2021.3.0 /Openmpi 4.1.1 nonl.f90: catastrophic error
Posted: Mon Aug 09, 2021 10:35 am
by marie-therese.huebsch
Hi hayak,
Thank you for reporting an issue. I would like to investigate it in detail, but more specific information is necessary. Please have a look at the
forum guidelines.
Best regards,
Marie-Therese
Re: Vasp 6.2.1 Intel Oneapi Compiler 2021.3.0 /Openmpi 4.1.1 nonl.f90: catastrophic error
Posted: Tue Aug 10, 2021 6:36 am
by hayak
Log files
Re: Vasp 6.2.1 Intel Oneapi Compiler 2021.3.0 /Openmpi 4.1.1 nonl.f90: catastrophic error
Posted: Tue Aug 10, 2021 7:30 am
by hayak
Output
Re: Vasp 6.2.1 Intel Oneapi Compiler 2021.3.0 /Openmpi 4.1.1 nonl.f90: catastrophic error
Posted: Tue Aug 10, 2021 9:38 am
by juergen.furthmueller
That's just an internal bug of a catastrophic compiler ... . You could only try another compiler (version).
Or is there simply some memory / stack size (limit) problem crashing the compiler?
Juergen
Re: Vasp 6.2.1 Intel Oneapi Compiler 2021.3.0 /Openmpi 4.1.1 nonl.f90: catastrophic error
Posted: Wed Aug 11, 2021 10:19 am
by hayak
Problem only with use -qopenmp parameter …
Re: Vasp 6.2.1 Intel Oneapi Compiler 2021.3.0 /Openmpi 4.1.1 nonl.f90: catastrophic error
Posted: Fri Aug 20, 2021 2:09 pm
by andreas.singraber
Hi hayak,
thank you for reporting this, I was able to reproduce the error with the Intel oneAPI compiler 2021.3. However, I was not using OpenMPI, your makefile.include also does not indicate that you used OpenMPI, are you sure about this (there is probably no connection to MPI)? Anyway, we will further investigate the origin of this.. in the meantime I can only recommend to stick to the 2021.2 version which worked well in my case. Can you try this out?
All the best,
Andreas