Link Search Menu Expand Document

Computing Resources

Condominium nodes on the Schooner supercomputer:

  • One dual-A100 GPU node (Nov. 2021)
    For deep learning on two A100 GPUs.
    Node ID: TBD.
    CPUs: dual 32-core AMD EPYC Rome 7452.
    GPUs: dual NVIDIA A100 40 GB.
    RAM: 256 GB.
    Disk: dual 960 GB SSD vSAS, RAID1.

  • Five Cascade Lake nodes (Jun. 2019)
    For parallel computing on the 200 CPU cores.
    Node ID: c657, c658, c659, c660 and c661.
    CPUs: dual 20-core Intel Xeon Gold 6230 Cascade Lake.
    RAM: 192 GB.
    Disk: 800 GB SSD SATA.

  • One Skylake node (Nov. 2018)
    For applications with large memory requirement.
    Node ID: c651.
    CPUs: dual 22-core Intel Xeon Gold 6152 Skylake.
    RAM: 384 GB
    Disk: dual 800 GB SSD SATA.

How to use the condominium nodes

Data storage:

  • local SSD space on compute nodes (/lwork)
  • 2 TB disk space on parallel file system (/work/omicsbio)
  • 10 TB tape archival space

How to manage the data and optimize the I/O

Computer servers

  • Thunder GPU server (Dec. 2020)
    For interactive deep learning on two RTX 3090 GPUs.
    URL: (SSH in OU intranet).
    CPUs: 24-core AMD Threadripper 3960X
    GPUs: dual RTX 3090 24 GB.
    RAM: 128 GB.
    Disk: dual 2TB GB SSD.
    How to use Thunder.

  • Virtual machines on the OU cloud
    Booming web server
    Milkyway database server