High Throughput Computing
While both High Performance Computing (HPC) and High Throughput Computing (HTC) jobs run parallel computations, HTC requires far less communication and synchronization between nodes. Parallel jobs commonly described as "embarrassingly parallel" (or "pleasantly parallel") are a perfect fit for HTC resources.
The processes comprising an HTC parallel computation are loosely-coupled, typically communicating with each other very little or not at all. HTC jobs can therefore be executed on physically distributed resources using grid-enabled technologies, such as XSEDE's Open Science Grid (OSG) and the University of Wisconsin-Madison's HTCondor project.
Typical HTC applications include Monte Carlo simulations and other parameter sweep applications, where the same program is run with varying inputs, resulting in hundreds or thousands of executions of the same program exemplifies a typical HTC job. HTC jobs typically run on a much-longer timescale than HPC jobs.
Available HTC resources
|Open Science Grid|
The Open Science Grid (OSG) advances science through open distributed computing. The OSG is a multi-disciplinary partnership to federate local, regional, community and national cyberinfrastructures to meet the researchers' high throughput computing needs.
Last update: January 20, 2015