Photo of the M3 cluster in the SMU Data Center.

M3

A high-performance computing (HPC) cluster is a shared-access supercomputer for computational research.

Sign In

M3

A high-performance computing (HPC) cluster is a shared-access supercomputer for computational research.

The M3 high-performance computing (HPC) cluster is a shared-access super computer for computational research. The Office of Information Technology (OIT) and the SMU Data Science Institute work together to provide the computational resources and tools required by SMU’s faculty, researchers, and students to accelerate advances in research innovation.

M3 features state of the art CPUs, networking, and storage technologies—making it one of the fastest HPC environments of any private university in Texas.

Resource Standard-Memory High-Memory
Nodes 170 8
Processors AMD EPYC 7763 AMD EPYC 7763
Frequency 2.45 GHz 2.45 GHz
CPUs/Node 2 2
Cores/Node 128 128
Memory/Node 512 GB 2 TB
Local Scratch/Node None 4.3 TB
Interconnect 200 Gb/s 200 Gb/s

OIT provides professional management of M3 with dedicated HPC staff and HPC-related research consulting through Research and Data Science Services. HPC technical support and consulting is provided to everyone in the SMU research community.

Support includes:

  • Getting started using M3
  • Computational research workflow optimization consulting
  • HPC programming performance and parallelization consulting
  • Data management consulting

Getting Started

If you are interested in using M3, you will first need to obtain an SMU HPC account. The account request form and more information about SMU HPC accounts is available on the HPC Accounts page. Student and other sponsored accounts are also available with a faculty sponsor approval.

Accessing the HPC Portal

  1. Click the Sign In button above to access the M3 HPC Portal.
  2. Sign in using your SMU ID and SMU password.

Please Note: Access to the HPC OnDemand web portal requires an existing SMU HPC account. If you log into the portal and don’t have an SMU HPC account, you will get an error message.