RHAPSODY on NSF and DOE Machines¶
This section provides machine-specific setup guides and verified examples for running RHAPSODY on NSF and DOE leadership computing facilities. These machines are periodically tested and verified against RHAPSODY and its pre-release changes.
Note
All of the tutorials in this section target multi-node runs using DragonExecutionBackend. For single-node runs, testing, or debugging, it is recommended to use ConcurrentExecutionBackend (ProcessPoolExecutor) instead.
Backend compatibility¶
ConcurrentExecutionBackend (designed for debug and tests single node only) works out of the box on every machine listed here — no machine-specific configuration is required. Install RHAPSODY and run.
DragonExecutionBackend (designed for large scale multi-node) requires machine-specific setup: each HPC system has its own software stack, MPI library, interconnect, and job launcher. The exact steps — module loading, virtual environment creation, transport configuration, and launch command — differ per machine and are documented in each section below.
Each guide covers environment setup, module loading, job allocation, and working examples.
Supported Machines¶
| Machine | Facility | Stack | Status |
|---|---|---|---|
| NCSA Delta | NSF / NCSA | Cray MPICH, SSH launcher | Verified |
| NERSC Perlmutter — Scalable Execution | DOE / NERSC | Cray libfabric + CUDA, SLURM launcher | Verified |
| NERSC Perlmutter — AI-HPC | DOE / NERSC | Cray libfabric + CUDA, SLURM launcher | Verified |
| PSC Bridges-2 | NSF / PSC | Standard SLURM, SSH launcher, TCP | Verified |
| Purdue Anvil | NSF / ACCESS / Purdue | Cray MPICH, SSH launcher (HSTA) | Verified |
More machines will be added as they are tested and validated.