MPI Cluster Overview¶
The MPI cluster enables jobs with tightly coupled parallel codes using Message Passing Interface APIs for distributing computation across multiple nodes, each with its own memory space.
Key Features¶
- Infiniband and Omni-Path networking
- Minimum of 2 full nodes per Job on mpi partition, minimum 1 full node on ndr partition
Specifications¶
| Partition | Host Architecture | Nodes | Cores/Node | Mem/Node | Mem/Core | Scratch | Network | Node Names |
|---|---|---|---|---|---|---|---|---|
mpi |
Intel Xeon Gold 6342 (Ice Lake) | 136 | 48 | 512 GB | 10.6 GB | 1.6 TB NVMe | HDR200; 10GbE | mpi-n[0-135] |
ndr |
AMD EPYC 9575F | 136 | 48 | 1.5 TB | 11.2 GB | 2.9 TB NVMe | NDR200; 10GbE | mpi-n[136-153] |
opa-high-mem |
Intel Xeon Gold 6132 (Skylake) | 36 | 28 | 192 GB | 6.8 GB | 500 TB SSD | OPA; 10GbE | opa-n[96-131] |