• Test bed
  • In service
  • Openly-available system for Prototyping and testing
  • Funded by STFC, DiRAC, ExCALIBUR
  • 1 nodes, with 4 AMD MI300A 128GB accelerators per node
  • Benchmarks (1) ▾
    • Memory bandwidth (BabelStream): 3648 GB/s
      • array_size: 134217728
      • iterations: 100
      • precision: FP64
  • Manufactured by AMD
  • Scheduler: Direct SSH
  • Interconnects:

COSMA MI300A

COSMA (The Compute Optimised System for Modelling and Analysis) is a High Performance Computing facility hosted at Durham University, operated by the Institute for Computational Cosmology on behalf of DiRAC.

The MI300A node is a GPU testbed within COSMA.

Node RAM CPU Access
ga008 500GB (unified) 96 cores (AMD EPYC) Direct SSH

The MI300A is an APU (Accelerated Processing Unit). The CPU and GPU are on the same silicon and share physical RAM. Therefore, memory copies between CPU and GPU are not necessary which can improve performance for many applications.

The MI300A node has 4x GPUs.

Documentation

Gaining access

Access requires a COSMA account, obtained via the DiRAC SAFE portal.

  1. Create a SAFE account with an institutional email.
  2. Upload an SSH public key on SAFE. If you do not have one, generate with ssh-keygen -t ed25519.
  3. Request a login account. This requires selecting a project, either:
    • Project do018 for AMD GPU testbed access.
    • A DiRAC project code for a given allocation (provided by a supervisor).
  4. Wait for the account to be approved by the project manager. Keep an eye on your email!
  5. Connect to COSMA via SSH: ssh username@login8.cosma.dur.ac.uk (Note: On first login you will be asked to change the password provided in your email)

Visit https://cosma.readthedocs.io/en/latest/account.html for more details. Contact cosma-support@durham.ac.uk for any questions.

Usage

Connect directly via SSH from a login node:

ssh ga008
rocm-smi
./gpu_program_to_run

The AMD ROCm stack can be loaded from the hipcc module on cosma. To load this module, check available versions with: module av hipcc Then load the desired version with: module load hipcc/VERSION

If your codebase is based on CUDA, you can convert this to HIP using the hipify script provided with ROCm.

Restrictions

  • /cosma5/ is not mounted on ga008. Copy binaries to your home directory first.
  • ga008 is a shared resource, please be mindful of other users when running codes. Use top to see running processes or users to see who else is logged in.