User-Available Nodes
Login Nodes*
1
Batch Nodes
30
Debug Nodes
2
Total Nodes
40
CPUs
CPU Architecture
Intel Xeon Silver 4215
Cores/Node
16
Total Cores
1,024
GPUs
GPU Architecture
NVIDIA Tesla
Total GPUs
128
GPUs per compute node
4
GPU global memory (GB)
32.00
Memory Total (GB)
12,256
CPU Memory/Node (GB)
383
Peak Performance
Peak TFLOPS (CPUs)
27.9
Peak TFLOPs (GPUs)
1,002.8
Peak TFLOPS (CPUs+GPUs)
1,030.60
Clock Speed (GHz)
2.5
OS
TOSS 4
Recommended location for parallel file space
Password Authentication
OTP, Kerberos
Compilers

See Compilers page