How to investigate stalled out frequency calculation

[Sibo Lin]

I have done numerical frequency calculations (direct mode) with a precompiled binary of NWChen 6.8 with no issue for dozens of molecules (~50 atoms), but when I try to run the same input file with a version of NWChem 7.0 that I compiled myself, the calculation stalls out by around doing the frequency of the 30th atom. I have checked that the compute node is at 100% cpu usage across 24 cores, and there is still memory available, but the frequency calculation does not update any files in my calculation directory, indefinitely.

Do you have any tips to diagnose why this is happening? I tried using “print debug” but the output file in the first minute was quite large, and I was afraid that it’d overwhelm storage space by the time the calculation got through 30 atoms. I will try to find a precompiled binary of NWChem 7.0 for Red Hat Enterprise Linux 6.7...to see if it was an issue with my compilation.

[Edoardo Aprà]

You can use the gdb debugger to attach the running processes and see where they are stuck.

You can find plenty of documentation on gdb's use on the web.

One page could the following one at section 20.2.2

https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/developer_guide/debugging-running-application

Could you please post the following file of your compilation to see if anything can be potentially problematic

$NWCHEM_TOP/src/tools/build/config.log

You can find EPEL6 nwchem 7.0.2 RPMs at

https://koji.fedoraproject.org/koji/buildinfo?buildID=1628753

[Sibo Lin]

I will try using one of the rpms you linked. That will be the simplest solution.  The logs from my own compilation is at https://raw.githubusercontent.com/sibo/chemScripts/master/config.log_debugCompiledNWChem7

--
You received this message because you are subscribed to a topic in the Google Groups "NWChem Forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/nwchem-forum/hTqB4JCt8gM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to nwchem-forum...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/26b42fd9-6768-494b-bf80-e01194c2326fo%40googlegroups.com.

[Edoardo Aprà]

The settings you used are correct but pretty inefficient.

I would suggest you to try the following

cd $NWCHEM_TOP/src/libext

unset USE_INTERNALBLAS

export BUILD_OPENBLAS=1

export BLAS_SIZE=8

export BUILD_SCALAPACK=1

export SCALAPACK_SIZE=8

make

cd ../tools

rm -rf build install

make

cd ../nwdft

make

cd ..

make link