DiaGrid + HUBzero: Power to the People
Category
Published on
Abstract
DiaGrid is a distributed computing resource based on a federated HTCondor pool with more than 50,000 processor cores today and has delivered nearly 100 million computational jobs to date. A new HUBzero-based web front end to DiaGrid, diagrid.org (aka DiaGrid HUB), is now available for researchers and students to access end-to-end scientific tools and workflows that require significant computation resources and, sometimes, complex orchestration. Diagrid.org brings large-scale distributed computing and high performance computing power to scientists and students in the form of powerful and yet easy-to-use scientific tools seamlessly integrated with and delivered to the appropriate computation and data resources on the back end. A growing list of popular research tools such as BLAST, R, and Gromacs are available now, with more tools in the pipeline. Faculty at Purdue and other partner institutions are also using diagrid.org to publish their research codes harnessing the vast number of processor cores and utilizing the Pegasus workflow engine now available through HUBzero.
This talk will describe the DiaGrid HUB and several scientific tools currently hosted there. It will also demonstrate one or two of these tools, as well as discuss future tools in the development pipeline. Feedback, suggestions and interest in collaboration are welcome and greatly appreciated.
Bio
Carol Song is a Senior Research Scientist and director of the Scientific Solutions group at the Rosen Center for Advanced Computing, Purdue University. She leads the DiaGrid project which delivers the power of high performance and high throughput computing resources to researchers and students via online scientific applications on HUBzero. She is the PI for Purdue's NSF TeraGrid Resource Partner and XSEDE projects, and currently is serving as the Chair of the XD Service Provider Forum. Carol received her Ph.D. in computer science from University of Illinois at Urbana-Champaign.
Cite this work
Researchers should cite this work as follows:
Submitter
Nikki Huang
Purdue University