Intrepid2: MonolothicExecutable Should Be Broken Up to Reduce Timeouts

[CamelliaDPG]

Bug Report

@intrepid2 The MonolothicExecutable times out reasonably often on various builds and platforms.

Description

As discussed by @cgcgcg in #14666, we would do well to break this test (which under the hood is divided into many individual tests) into smaller tests. Ideally, the new tests would have a maximum of about 10% the runtime of the current MonolothicExecutable.

For developer convenience, we might maintain the MonolothicExecutable as a target that is not by default included in the tests -- this allows us to use e.g. glob matches to run all the HGRAD tests within MonolothicExecutable.

[cgcgcg]

@CamelliaDPG You could take an approach where you keep the same ADD_EXECUTABLE for MonolithicExecutable but have different ADD_TEST declarations with filters for the unit tests.

[CamelliaDPG]

Running a debug build on my workstation, the monolithic executable tests take about 32 seconds.

        Start 167: Intrepid2_unit-test_MonolithicExecutable_Intrepid2_Tests_MPI_1
167/169 Test #167: Intrepid2_unit-test_MonolithicExecutable_Intrepid2_Tests_MPI_1 ........................................   Passed   31.87 sec

So I'm targetting about 3 seconds each for each of the new tests.

[CamelliaDPG]

@CamelliaDPG You could take an approach where you keep the same ADD_EXECUTABLE for MonolithicExecutable but have different ADD_TEST declarations with filters for the unit tests.

Thanks, @cgcgcg, for this suggestion! It was a much cleaner way to do it than what I'd been thinking of. The PR is #14784.