Search open search form

Resources

Caltech HPC Resources

  • The Caltech Supercomputing Lab offers a total of 12,840 CPU cores, 200 GPUs, and over 100 TB of memory across 337 compute nodes.
  • Compute resources are supported by approximately 2.5 petabytes of storage.
  • All nodes run CentOS Linux 7.
  • Each GPU node hosts 4 Nvidia P100 GPU's each containing 16GB of memory. One node has 2 p100s and 2 v100s

Interactive login nodes:

Hostname Accessibility Nodes Cores/Node CPU Type Memory/Node Memory Speed
login.hpc.caltech.edu Caltech Only 2 32 16-core Intel Skylake (2.1  GHz) 192 GB TBD

Compute nodes:

Type Nodes Total Cores Total Memory Cores/Node CPU Type Memory/Node Memory Speed Walltime Limit Infiniband GPUs/Node
Compute 179 5728 34,368 GB 32 16-core Intel Skylake (2.1 GHz) 192 GB 2666 MT/s
7 Days EDR
Large Memory 11 352 4,224 GB 32 16-core Intel Skylake (2.1 GHz) 384 GB 2666 MT/s
7 Days
EDR
Expansion Compute 85 4760 32,810 GB 56 28-core Intel Cascadelake 8276 (2.20 GHz) 386 GB 2933 MT/s 7 Days EDR
Expansion Big Mem Compute 9 504 3 TB 56 28-core Intel Cascadelake 8276 (2.20 GHz) 1.5 TB 2933 MT/s 7 Days EDR
Expansion Big Mem 4-way Compute 2 192 3 TB 96 24-core Intel Cascadelake 8276 (2.40 GHz) 1.5 TB 2933 MT/s 7 Days EDR
GPU 50 1,400 12,800
 GB
28 14-core Intel Broadwell (2.4 GHz) 256 GB 2400 MT/s 7 Days
EDR 4