{cae}[dummy/dummy] FDTD_Solutions v8.20.1731

[migueldiascosta]

(created using eb --new-pr)

requires easybuilders/easybuild-easyblocks#1307

[migueldiascosta]

Test report by @migueldiascosta
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in this PR)
grc-cluster1 - Linux centos 6.10, Intel(R) Xeon(R) CPU E5-2640 0 @ 2.50GHz, Python 2.7.14
See https://gist.github.com/d22b59cfa0437fa004771d8a5b33bb15 for a full test report.

[migueldiascosta]

@akesandgren this comes packaged with mpich2 libraries, but would need impi or OpenMPI to be loaded if those are to be used, should I add separate easyconfig files with the respective dependencies? how have you been using it?

[akesandgren]

We used dummy tc and didn't add any MPI stuff. But there are a couple of patches that would probably be useful. See #5502

And easybuilders/easybuild-easyblocks#1307 is also useful (still needs a review and merge)

[migueldiascosta]

hm, I missed #5502, I'll have a look

(easybuilders/easybuild-easyblocks#1307 is precisely what this was meant to test :) )

[migueldiascosta]

added one patch from #5502, the other two don't seem to apply any more

[migueldiascosta]

Test report by @migueldiascosta
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in this PR)
grc-cluster1 - Linux centos 6.10, Intel(R) Xeon(R) CPU E5-2640 0 @ 2.50GHz, Python 2.7.14
See https://gist.github.com/28346923333374a06151f310d6df6f63 for a full test report.

[boegel]

@akesandgren Maybe also submit a test report for this (and then merge if it's good to go)?

[akesandgren]

We only have 8.16.982 so can't test this one.

[akesandgren]

@migueldiascosta doesn't it still hardcode paths to /opt in all the scripts and the python script?

[migueldiascosta]

@akesandgren their python script has been fixed. The bundled MPI is now mpich2/nemesis and it only ships mpiexec binaries and the libraries, no mpirun scripts. The fdtd-run-local.sh now uses this mpiexec.

For other MPI libraries, it seems we it is now needed to use their own launchers, hence the question I asked at the beginning of this PR about loading ipmi or OpenMPI.

There is indeed one additional hardcoded path that you fixed in your patch, the one in templates/fdtd-pbs-template.sh. Not only is it still hardcoded, it's hardcoded to a now non-existing path, I'll fix that.

[migueldiascosta]

Test report by @migueldiascosta
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in this PR)
grc-cluster1 - Linux centos 6.10, Intel(R) Xeon(R) CPU E5-2640 0 @ 2.50GHz, Python 2.7.14
See https://gist.github.com/e3a8e0c90f5e249920feebdb1b15312f for a full test report.

[akesandgren]

LGTM

[akesandgren]

Going in, thanks @migueldiascosta!