Stampede3 node architecture
Stampede3 is operated by Texas Advanced Computing Center in United States. This profile covers Sapphire Rapids and Skylake CPU partitions. Specifications are listed below; anything the operator has not published is itemised in the documented-fields list.
Cluster · Sapphire Rapids HBM
- Operator / site
- Texas Advanced Computing Center
- Country
- United States
- Profile scope
- Sapphire Rapids and Skylake CPU partitions
- Data checked
- 14 September 2026
- Operating system
- Rocky Linux
- Network topology
- Fat-tree (100 Gb/s Omni-Path)
Sapphire Rapids HBM
- Compute nodes of this type
- 560
- CPU model
- Intel Xeon CPU Max 9480 · 56 cores
- CPU sockets per node
- 2
- Cores per socket
- 56
- Physical CPU cores
- 112
- Host / unified memory
- 128 GB HBM2e
- Interface speed
- 100 Gb/s Omni-Path
- Local storage
- 150 GB /tmp
Skylake
- Compute nodes of this type
- 1060
- CPU model
- Intel Xeon Platinum 8160 · 24 cores
- CPU sockets per node
- 2
- Cores per socket
- 24
- Physical CPU cores
- 48
- Host / unified memory
- 192 GB DDR4
- Interface speed
- 100 Gb/s Omni-Path
- Local storage
- 90 GB /tmp
Scope and uncertainty
Sapphire Rapids and Skylake CPU partitions only. Ice Lake and accelerator partitions are outside this profile; totals are not the whole installation.
Documented fields: 18 / 33
Every field this catalog checks. Fields listed above have a value; the remainder are shown here. A field marked absent is not publicly available in the sources we could verify, which does not mean the hardware is missing. Fields that cannot apply, such as benchmark results for an unranked system, are excluded from the total.
Cluster — 2 / 3
- ✕Manufacturer
- –Ranked core count
- –Theoretical peak (PFlop/s)
- –Measured power (kW)
- –Energy efficiency (GFlop/s per watt)
- ✓Operating system
- ✓Network topology
Sapphire Rapids HBM — 8 / 15
- ✕System / node model
- ✓Compute nodes of this type
- ✓CPU model
- ✓CPU sockets per node
- ✓Cores per socket
- ✓Physical CPU cores
- ✕CPU note
- –Accelerator model
- –Physical accelerators per node
- ✓Host / unified memory
- –Accelerator memory
- –Accelerator ↔ accelerator link
- –On-node fabric
- –CPU ↔ accelerator link
- ✕Network interfaces per node
- ✕Interface model
- ✓Interface speed
- ✕Interface-to-accelerator topology
- –Aggregate scale-out per node
- –GPUDirect RDMA
- ✕Management / storage networking
- ✓Local storage
- –NVMe data disks
- –Max remote data disks
- –Uncached disk IOPS
- –Uncached disk throughput (MBps)
- ✕Form factor
Skylake — 8 / 15
- ✕System / node model
- ✓Compute nodes of this type
- ✓CPU model
- ✓CPU sockets per node
- ✓Cores per socket
- ✓Physical CPU cores
- ✕CPU note
- –Accelerator model
- –Physical accelerators per node
- ✓Host / unified memory
- –Accelerator memory
- –Accelerator ↔ accelerator link
- –On-node fabric
- –CPU ↔ accelerator link
- ✕Network interfaces per node
- ✕Interface model
- ✓Interface speed
- ✕Interface-to-accelerator topology
- –Aggregate scale-out per node
- –GPUDirect RDMA
- ✕Management / storage networking
- ✓Local storage
- –NVMe data disks
- –Max remote data disks
- –Uncached disk IOPS
- –Uncached disk throughput (MBps)
- ✕Form factor
Sources
Browse all cluster architecture pages →
CLUSTERATLAS