Expanse node architecture
Expanse is operated by San Diego Supercomputer Center in United States. This profile covers Standard CPU and GPU partitions.
Cluster interconnect: InfiniBand HDR. Topology: Not verified.
Standard CPU
- Node count
- 728
- CPU
- AMD EPYC 7742 · 64 cores
- CPUs per node
- 2
- Accelerator
- No accelerator
- Physical accelerators per node
- 0
- Host / unified memory
- 256 GB DRAM
- Accelerator memory
- Not verified
- Local accelerator links
- Not applicable
- CPU-to-accelerator link
- Not verified
- Storage
- 1 TB NVMe
V100 GPU
- Node count
- 52
- CPU
- Intel Xeon Gold 6248 · 20 cores
- CPUs per node
- 2
- Accelerator
- NVIDIA V100
- Physical accelerators per node
- 4
- Host / unified memory
- 384 GB DRAM
- Accelerator memory
- 32 GB per GPU
- Local accelerator links
- NVLink
- CPU-to-accelerator link
- Not verified
- Storage
- 1.6 TB NVMe
Scope and uncertainty
Standard CPU and GPU partitions only. Four large-memory nodes are excluded from this profile pending detailed verification.
Not verified means a detail has not been confirmed in this catalog. Diagrams are schematic; node indices are illustrative, not physical hostnames. Data checked 14 September 2026.
Sources
Browse all cluster architecture pages →
CLUSTER