Do You Need Discovery?
Discovery is a supercomputer located at New Mexico State University, available to researchers around NMSU for free. Discovery has a high level of computing performance compared to a general-purpose computer. It can be used for a wide range of computationally intensive tasks in various fields. One can use it for mathematical calculations, fluid dynamics calculations, pattern matching, graph analysis, data collection, nuclear energy research, astrophysics simulations, bioinformatics, and more. Discovery runs a Linux-based operating system, like most supercomputers today.
Discovery cluster is a collection of computers, or nodes, that communicate using InfiniBand, making it an ideal location to scale computational analysis from your personal computer. The high-performance computing (HPC) cluster is free to use for current students, faculty, and staff for both research, and teaching. Discovery is equipped with Lenovo nx360, SR650, SR670, SR850, Dell R440, and R740 hardware, and it runs Slurm scheduling manager.
Discovery can perform more operations than your personal computer, and much faster. It currently contains 1536 cores working together, with over 1 Petabyte of storage, and it will continue to grow with new nodes being added. Aside from being able to get research results faster, there is help available right on campus. The High-Performance Computing (HPC) team can help you with all "Discovery" related tasks, including adapting your code, job submission, and how to use the scheduler.
However, if you need more resources than Discovery has, you can try using ACCESS. To learn more, you can contact go to https://access-ci.org/ or contact NMSU Campus Champions.
Local Desktop |
Discovery |
ACCESS |
|
# of cores |
4 |
1000s |
1000s |
Memory (GB) |
8 |
100s |
1000s |
Storage (TB) |
1 |
100s |
1000s |
Time Limit |
none |
medium-long |
short-medium |
Help |
limited |
local |
limited |
Discovery’s Hardware Configuration
The team maintains one HPC cluster currently and the below table shows the hardware configuration of the HPC cluster.
HPC Cluster Configuration | Values |
---|---|
OS |
Red Hat Enterprise Linux release 8 (RHEL 8) |
Cores |
1536 |
Scheduler |
Slurm 22.05.11 |
Interconnect |
InfiniBand HDR (100Gbps) |
Total Nodes |
54 (38 CPU nodes, 16 GPU nodes) |