Add E3SM-Unified support to test suite

[xylar]

This merge adds a bash script for testing E3SM-Unified. Various changes to the setup script and templates were needed to make this work.

[xylar]

Testing

I ran the test suite using E3SM-Unified v1.7.0rc1 on Anvil and Chrysalis. Results are here:
https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.xylar/analysis_testing/anvil/test_e3sm_unified_1.7.0rc1_anvil/
and here:
https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.xylar/analysis_testing/chrysalis/test_e3sm_unified_1.7.0rc1_chrysalis/

[xylar]

Cori and Compy

The suite also ran successfully on Compy:
https://compy-dtn.pnl.gov/asay932/analysis_testing/compy/test_e3sm_unified_1.7.0rc1_compy/
and is running Cori-Haswell (mostly unfinished because of long queue times):
https://portal.nersc.gov/project/e3sm/xylar/analysis_testing/cori-haswell/test_e3sm_unified_1.7.0rc1_cori-haswell/

The results on Compy are showing some remapping errors that we saw in testing the previous MPAS-Analysis in the previous E3SM-Unified (1.6.0):

I will need to remind myself what the cause and solution was, and see what can be done here as well. As that behavior is unrelated to this PR, I will address it elsewhere

[xylar]

Update on the bad remapping. The last time this happened, it was on Chrysalis. The solution was to build E3SM-Unified with Gnu compilers rather than Intel. That is currently not an option on Compy but I will add it tomorrow, and then build the next test environment for E3SM-Unified with Gnu.

[xylar]

I will merge this PR as soon as the analysis finishes successfully on Cori-Haswell.