Fix deadlock and message file retention on Frontier.

[joaander]

Description

• Remove simulation.walltime evaluations from inside rank == 0 checks.
• Fix other miscellaneous issues on Frontier.

Motivation and context

• communicator.walltime calls MPI_Bcast and therefore must be called collectively. I opted to remove the walltime message rather than call communicator.walltime collectively and print the result.
• The message_file was overwritten on job restart. Tag the file with the slurm job id to keep old logs.
• Prevent the new patchy particle test from running when HOOMD is built without LLVM.

How has this been tested?

For unknown reasons, this does not always deadlock as I would expect. It only deadlocks in very rare circumstances in job submissions on OLCF Frontier. I have tested the create_initial_state operations on Frontier and they work now.

TODO:

• Test full workflow.

Checklist:

• I have reviewed the Contributor Guidelines.
• I agree with the terms of the HOOMD-blue Contributor Agreement.
• My name is on the list of contributors (CONTRIBUTORS.md) in the pull request source branch.

[joaander]

The entire workflow runs on frontier.