Skip to main content

IO Benchmarking and Testing in Support of the Global Cloud Resolving Model and Data Subsetting and Analysis


EMSL Project ID
24094

Abstract

This proposal is for a development/test account for the EMSL MPP2 cluster,
Lustre file system, and visualization server in support of the recently
funded SciDAC proposal Community Access to Global Cloud Resolving Model
Data and Analyses. This work is being performed in conjunction with
the SciDAC project Design and Testing of a Global Cloud Resolving Model,
led by Professor Randall at Colorado State University.

This project is dedicated to providing efficient, flexible access to
logical subsets of Global Cloud Resolving Model (GCRM) output as well
as selected derived products and visualizations, thus enabling a range
of analyses by the broader climate research community. Achieving this
first requires substantial effort in determining the best way to perform
IO such that IO does not significantly block computation, while ensuring
that the layout and partitioning of the data allows for efficient data
access for analysis. While, the GCRM is ultimately targeted at the DOE
petascale facilities at ORNL, access to EMSL's MPP2 and the Lustre file
system is required during the development phase of this project.

Project Details

Project type
Limited Scope
Start Date
2007-03-01
End Date
2007-05-01
Status
Closed

Team

Principal Investigator

Karen Schuchardt
Institution
Pacific Northwest National Laboratory

Team Members

Jeffrey Daily
Institution
Pacific Northwest National Laboratory

Todd Elsethagen
Institution
Pacific Northwest National Laboratory

Annette Koontz
Institution
Pacific Northwest National Laboratory

Bruce Palmer
Institution
Pacific Northwest National Laboratory

Related Publications

Schuchardt KL, BJ Palmer, J Daily, TO Elsethagen, and AS Koontz. 2007. "IO strategies and data services for petascale data sets from a global cloud resolving model." Journal of Physics: Conference Series 78:Art. No. 012089.