CAMSA is a Dell cluster with a rich accelerator testbed consisting of Intel MAX GPUs (Graphic Processing Units), Intel FPGAs (Field Programmable Gate Arrays), NVIDIA and Graphcore IPUs. The CAMSA cluster consists of compute nodes using a mix of the following processors.
- Power Edge R650 Head Node x2: Intel Xeon Gold 6330 512 RDIMM 2TB
- Power Edge R650 Login Node x2: Intel Xeon Gold 5118 256 RDIMM 2TB
- Power Edge R420 Compute Node x5: Intel Xeon Gold 622XR 512 RDIMM 2TB
- Power Edge R750 GPU Node: Intel Xeon Gold 6330 512 RDIMM 2TB
- Power Edge R650 NFS Storage Node x5: Intel Xeon Gold 6330 512 RDIMM 2TB
- Power Edge C6420 x10: Intel Xeon Gold 6226R 2.9G, 16C/32T, 10.4GT/s, 22 M Cache, Turbo, HT (150W) DDR4-2933
CAMSA Overview
- Slurm‑managed HPC cluster (with backfill scheduling)
- 20 total compute nodes
- 2 login nodes
- 18 CPU compute nodes (32 cores per node; 600+ cores total)
- 4 GPU compute nodes
- 6 × NVIDIA A100‑PCIE‑40GB GPUs (2 per node)
- Separate CPU and GPU job partitions
GPU Capabilities
- NVIDIA A100 GPUs with 40 GB HBM2 memory
- Tensor Core acceleration
- CUDA 12.x support
Use Cases
- Modeling and simulation
- Machine learning and AI
- Data‑intensive analytics
- Faculty‑student research
- NSF‑funded research projects