Cluster computing
Cluster computing involves sets of computing nodes that are networked together to perform common computing tasks. Computing clusters can range from clusters built for single projects to very large supercomputing center clusters which are used for the largest computational research projects in the world.
High Performance Computing (HPC)
HPC uses all available computing resources to complete many tasks all at once, in a short period of time. HPC generally deals with one large problem for which single processor performance is important. HPC is usually a good fit when:
- your datasets are too big for your personal machine
- you have computational problems can't be easily re-started or re-run, and for which reliability is important (e.g., if one part fails, the entire calculation fails)
- you have computing tasks that are dependent on intermediate results
Learn more about High Performance Computing at NC State.
High Throughput Computing (HTC)
In contrast to HPC, HTC is focused on computing environments that can deliver large amounts of processing capacity over long periods of time. HTC workflows decompose and parallelize computing tasks and run them many times over many computers. Some hallmarks of research that can likely be run in a high-throughout way include:
- workflows that process many samples, images, models, parameters, etc.
- many scientific simulations and statistical analyses or modeling
- tasks that run for a long time
Explore developing HTC workloads through the Open Science Grid.
National Compute Resources
If your computational workloads are larger than what can be accommodated at NC State or you have extremely specialized computational needs, you may be able to benefit from one of the many computing resources sponsored by national funding agencies and maintained by various institutions and consortia around the country.
Learn more about nationally available compute resources.
You can also contact us for a consultation.