enhance PETSc easyblock to run tests in parallel

[casparvl]

The PETSc test suite can run in parallel by passing a -j <n> (see https://petsc4py.readthedocs.io/en/stable/manual/tests/) but the previous EasyBlock would just run things serially.

I've tested with 64 tasks (on a 128 core node), which speeds up the test suite from about 3.5 hours to about 30 mins. On another 16-core node, with 16 tasks, it was even down to 17 minutes (probably faster filesystem)

[akesandgren]

This should probably be synced with easybuilders/easybuild-framework#3842 although I'm not sure if it has an impact or not.

[boegel]

Test report by @boegel

Overview of tested easyconfigs (in order)

• SUCCESS PETSc-3.12.4-foss-2019b-Python-3.7.4.eb
• SUCCESS PETSc-3.12.4-foss-2020a-Python-3.8.2.eb
• SUCCESS PETSc-3.12.4-intel-2019b-Python-2.7.16.eb
• SUCCESS PETSc-3.12.4-intel-2019b-Python-3.7.4.eb
• SUCCESS PETSc-3.12.4-intel-2020a-Python-3.8.2.eb
• SUCCESS PETSc-3.14.4-foss-2020b.eb
• SUCCESS PETSc-3.14.4-intel-2020b.eb

Build succeeded for 7 out of 7 (7 easyconfigs in total)
node2618.swalot.os - Linux centos linux 7.9.2009, x86_64, Intel(R) Xeon(R) CPU E5-2660 v3 @ 2.60GHz (haswell), Python 3.6.8
See https://gist.github.com/c4d2263f3cbd928bf849925ddb13058c for a full test report.

[boegel]

Test report by @boegel

Overview of tested easyconfigs (in order)

• SUCCESS PETSc-3.12.4-foss-2019b-Python-3.7.4.eb
• SUCCESS PETSc-3.14.4-foss-2020b.eb
• SUCCESS PETSc-3.15.1-foss-2021a.eb

Build succeeded for 3 out of 3 (3 easyconfigs in total)
node3553.doduo.os - Linux RHEL 8.2, x86_64, AMD EPYC 7552 48-Core Processor (zen2), Python 3.6.8
See https://gist.github.com/91b0f3e6729fa6103573bd25dc41a601 for a full test report.

[casparvl]

@akesandgren : I'd probably not make this PR dependend on what is decided in easybuilders/easybuild-framework#3842 , it would delay integration of this block until a decision has been taken in that other PR. Also: the notion in that PR (for now) is not to break EasyBlocks that use ec[parallel] currently, and if we in the end decide to deviate from that, we'll have to find & change all EasyBlocks in which this is used anyway. We might as well update the PETSc EasyBlock then - because what is one more? :)

[boegel]

Test report by @boegel

Overview of tested easyconfigs (in order)

• SUCCESS PETSc-3.9.3-foss-2018a.eb
• SUCCESS PETSc-3.9.3-intel-2018a.eb
• SUCCESS PETSc-3.11.1-foss-2019a-Python-3.7.2.eb
• SUCCESS PETSc-3.11.1-intel-2019a-Python-3.7.2.eb

Build succeeded for 4 out of 4 (4 easyconfigs in total)
node2602.swalot.os - Linux centos linux 7.9.2009, x86_64, Intel(R) Xeon(R) CPU E5-2660 v3 @ 2.60GHz (haswell), Python 3.6.8
See https://gist.github.com/ced64516001c4fc224d22cd494771bce for a full test report.

[boegel]

lgtm