SMU launches powerful new supercomputing research system with NVIDIA. The NVIDIA DGX SuperPOD will fuel transformational high-performance computing ecosystem for SMU and North Texas.
The mission of the Center for Research Computing at Southern Methodist University is to support research and education enabled by high-performance computing technology. We are motivated by the fact that computer simulation and data analysis has become an essential facilitator of discovery in most disciplines, and that advances in computing, networking, and data storage technologies are likely to accelerate this trend.
The primary purposes of the Center are to: (i.) provide a state-of-the-art research computing infrastructure for SMU faculty and students; (ii.) provide training and support to faculty and students in the use of the Center's resources and external facilities, including awarding certificates in high-performance computing based on a completion of assignments associated with center's workshops; (iii.) provide the mechanism for faculty governance of advanced research computing and related educational activities ; (iv.) stimulate multidisciplinary research involving computation; (v.) support and develop educational programs involving computation; (vi.) publicize faculty and student research involving computation; and (vii.) engage with local government and industry on relevant research projects. An annual report of these activities is available here.
ManeFrame II (M2)
SMU’s high-performance compute cluster provides vast computational capacity and performance for researchers at SMU. The cluster features state of the art CPUs, accelerators, networking technologies, significant amounts of memory per node, interactive GPU-accelerated remote desktops, and web-based access via hpc.smu.edu.
ManeFrame II runs the CentOS 7 operating system, the SLURM resource scheduler, and the Lmod environment module system. Additionally, various development tool chains are be available including the GCC, Intel, and PGI compiler suites. Optimized high-level programming environments such as MATLAB, Python, and R are also available in addition to the domain specific software packages that SMU researchers depend on for their work.
More information on ManeFrame II and use can be found here.
| ManeFrame I (retired) | ManeFrame II (2017) | ManeFrame II (2019) | |
|---|---|---|---|
| Computational Ability | 104 TFLOPS | 630 TFLOPS | 870 TFLOPS |
| Number of Nodes | 1,104 | 349 | 354 |
| Intel CPU Cores (AVX2) | 8,832 | 11,088 | 11,276 |
| Total Accelerator Cores | 0 | 132,608 | 275,968 |
| Total Memory | 29.2 TB (29,856 GB) | 116.5 TB (119,336 GB) | 120 TB (122,880 GB) |
| Node Interconnect Bandwidth | 20 Gb/s | 100 Gb/s | 100 Gb/s |
| Scratch Space | 1.4 PB (1,229 TB) | 1.4 PB (1,434 TB) | 2.8 PB (2,867 TB) |
| Archive Capabilities | No | Yes | Yes |
| Operating System | Scientific Linux 6 | CentOS 7 | CentOS 7 |
NVIDIA DGX SuperPOD Advantage Specifications
The SMU SuperPOD is a high-performance computing (HPC) cluster, specifically tailored to meet the demands of cutting-edge research. This shared resource machine consists of 20 NVIDIA DGX A100 nodes, each with 8 advanced and powerful graphical processing units (GPUs) to accelerate calculations and train AI models. The SMU Office of Information Technology (OIT) and the Center for Research Computing (CRC) jointly manage and provide both access and support for this top of the line machine.
If you are interested in learning more about how this machine can help your research, please reach out to the OIT Research and Data Science Services group viahelp@smu.eduwith HPC on the subject line.
| Computational Ability | 1,644 TFLOPS |
| Number of Nodes | 20 |
| CPU Cores | 2,560 |
| GPU Accelerator Cores | 1,392,640 |
| Total Memory | 52.5 TB |
| Node Interconnect Bandwidth | 10 - 200 Gb/s Infiniband Connections Per Node |
| Work Storage | 768 TB (Shared) |
| Scratch Storage | 750 TB (Raw) |
| Archival Storage | N/A |
| Operating System | Ubuntu 20.04 |
Request an Account
Accounts for Individuals with an SMU ID
- SMU faculty/staff can request an account by filling out the New Account Form.
- If you are a student or postdoc please ask your supervisor, sponsor, or adviser to request an account via the New Account Form.
Accounts for Individuals without an SMU ID or Sponsored ID
- The sponsor must request an SMU ID or Sponsored ID and have that account activated before sponsor can request ManeFrame II (M2) access for the individual. Please submit an Account Request via the Online Service Center.
- Once the individual's SMU ID or Sponsored ID has been created and activated the sponsor can request an ManeFrame II (M2) account for the individual via the New Account Form.
Documentation
Here you can can find information on how to effectively use SMU’s HPC resources. Topics covered in the documentation include:
For questions about using resources or setting up accounts please email the SMU HPC Admins with "HPC" in the subject line.
CRC HPC Fall 2022 Workshop Series
The Fall 2022 OIT, CRC, & DSI workshop series provides a hands-on experience that will guide researchers from the basics of using SMU’s supercomputing resources to advanced parallelization and application specific usage. The topics will cover information useful for researchers to quickly begin to use the advanced compute capabilities provided with the cluster. New users are encouraged to take advantage of the introductory “Introduction to HPC” workshop that will be given once monthly during the semester.
Fall 2022 OIT, CRC, & DSI Workshop Series - Open to All!
|
Date |
Time |
Workshop |
Series |
Location |
|
8/23 |
10:00-12:00 |
Introduction to HPC |
1/1 |
Ford 124 |
|
8/30 |
12:30-2:30 |
Data Mining Using Rapid Miner |
1/1 |
FLE 109 |
|
9/2 |
10:00-12:00 |
C++ Programming Workshop |
1/6 |
Ford 124 |
|
9/6 |
10:00-12:00 |
Introduction to HPC |
1/1 |
Ford 124 |
|
9/7 |
12:30-2:30 |
Machine Learning in R |
1/2 |
FLE 109 |
|
9/9 |
12:30-2:30 |
Machine Learning in R |
2/2 |
FLE 110 |
|
9/14 |
12:30-2:30 |
Text Mining in R |
1/1 |
FLE 109 |
|
9/16 |
10:00-12:00 |
C++ Programming Workshop |
2/6 |
Ford 124 |
|
9/20 |
10:00-12:00 |
Parallel C++ STL |
1/1 |
Ford 124 |
|
9/21 |
12:30-2:30 |
Machine Learning in Python |
1/2 |
FLE 109 |
|
9/23 |
12:30-2:30 |
Machine Learning in Python |
2/2 |
FLE 110 |
|
9/27 |
10:00-12:00 |
Parallel C++ APIs |
1/1 |
Ford 124 |
|
9/30 |
10:00-12:00 |
C++ Programming Workshop |
3/6 |
Ford 124 |
|
10/4 |
10:00-12:00 |
Introduction to HPC |
1/1 |
Ford 124 |
|
10/5 |
12:30-2:30 |
Deep Learning with Python |
1/1 |
FLE 109 |
|
10/14 |
10:00-12:00 |
C++ Programming Workshop |
4/6 |
Ford 124 |
|
10/18 |
10:00-12:00 |
Using MPI |
1/1 |
Ford 124 |
|
10/25 |
10:00-12:00 |
Bridging C++ and Python |
1/1 |
Ford 124 |
|
10/28 |
10:00-12:00 |
C++ Programming Workshop |
5/6 |
Ford 124 |
|
11/1 |
10:00-12:00 |
Introduction to HPC |
1/1 |
Ford 124 |
|
11/8 |
10:00-12:00 |
Bridging C++ and R |
1/1 |
Ford 124 |
|
11/11 |
10:00-12:00 |
C++ Programming Workshop |
6/6 |
Ford 124 |
Report HPC Usage
Faculty are encouraged to report their usage of SMU’s HPC facilities here. The data will be used to assess and document usage of these resources.Acknowledgement
If you wish to acknowledge the use of M2 in any publication we suggest the following: "Computational resources for this research were provided by SMU's Center for Research Computing."