Michigan Tech's shared high-performance computing infrastructure, Superior, is available to all researchers. It has the following computing and storage components:
- Generation 1.0 (acquired between 2013/06 - 2015/10)
- 92 CPU compute nodes - each having 16 CPU cores (Intel Xeon E5-2670 2.60 GHz) and 64 GB RAM - providing 30 TFLOPS
- 4 CPU compute nodes - each having 24 CPU cores (Intel Xeon E4-2680 2.50 GHz) and 256 GB RAM - providing 2 TFLOPS
- 5 GPU compute nodes - each having 16 CPU cores (Intel Xeon E5-2670 2.60 GHz), 64 GB RAM and 4 NVIDIA Tesla M2090 GPUs - providing 13 TFLOPS
- 3 storage nodes each with 32 TB usable space
- Generation 2.0 (acquired between 2017/06 - 2018/08)
- 85 CPU compute nodes - each having 32 CPU cores (Intel Xeon E5-2683 2.10 GHz) and 256 GB RAM - providing 91 TFLOPS
Portage is another shared high-performance computing infrastructure and a miniature version of Superior. Intended primarily for testing, educational (course work and/or senior design projects) and gateway/preliminary research projects involving non-confidential/non-sponsored data, Portage has 3 TFLOPS of CPU and 2 TFLOPS of GPU computing capacity with hardware identical to Superior's generation 1.0.
Superior and Portage have a Gigabit ethernet back-end network that serves the administrative needs, and a 56 Gb/s InfiniBand network that serves the computing needs. They are available for all researchers at Michigan Tech via a brief proposal, very much similar to that of NSF XSEDE.
Researchers can get help with hardware specification and acquisition, software acquisition and licensing, compilation, installation, integration with the queuing system, running benchmarks, developing computational workflows, and necessary end-user training.
Michigan Tech's Director of Research Computing, Dr. Gowtham, serves as the NSF XSEDE Campus Champion. He has access to and trial allocations in several supercomputers around the country and will help researchers find an optimal external resource, if necessary.