Parametric sweep provides a straightforward development path for solving delightfully parallel problems on a cluster (sometimes referred to as "embarrassingly parallel" problems, which have no data interdependencies or a shared state that would preclude linear scaling through parallelization). One such problem is the calculation of prime numbers over a large range. Parametric sweep applications run multiple instances of the same program on different sets of input data stored in a series of indexed storage items, such as files on a disk or rows in a database table. Each instance of a parametric sweep application runs as a separate task, and many such tasks can execute concurrently, depending on the amount of available cluster resources.
Hands-On Labs
CPU Spinner
This sample demonstrates how to run an HPC parametric sweep application on Windows Azure compute nodes.
NCBI BLAST
This sample demonstrates how to run a parametric sweep application that matches inputted nucleotide sequences to the human genome. The sample also demonstrates how to submit a job using the HPC Job Scheduler’s REST API and how to upload and download files from Windows Azure blob and table storage.
Image Rendering
This sample demonstrates how to run a parametric sweep application that renders images according to instruction files and how to upload and download files from Windows Azure blob storage.
Image Rendering v2
This sample demonstrates how to run a parametric sweep application that renders images according to instruction files, how to upload and download files from Windows Azure blob storage, and how to call the HPC Job Scheduler Web service with the REST API