Before running these tests you should check that the unit tests pass, see here.
The parallel unit tests can be run with
source run_test.sh > outfile.log
This will run the same tests multiple times, with different numbers of MPI ranks, and write a large number of PASS or FAIL statements to outfile.log. We do not expect any FAILs, but the log will contain a few other statements (e.g. multidomain summary information).
These tests check for accuracy, not parallel performance. We do not expect speed-ups proportional to the number of ranks, because the etst code is not written to scale well. But it might detect problems with your MPI setup.
To quickly count the FAIL statements I suggest
cat outfile.log | grep FAIL | wc -w
which should return 0 if there are no failures. The same approach can be applied to PASS statements
cat outfile.log | grep FAIL | wc -w
and at the time of writing returns 4215.