Skip to main content
SHARE
Research Highlight

A Co-design Framework for Online Data Analysis and Reduction

The Cheetah-Savanna workflow suite. Cheetah is the user interface to compose campaigns using Python. Savanna is the execution engine to run a campaign on a target HPC machine. CSMD ORNL Computer Science and Mathematics Oak Ridge
The Cheetah-Savanna workflow suite. Cheetah is the user interface to compose campaigns using Python. Savanna is the execution engine to run a campaign on a target HPC machine.

A group of researchers from ORNL, ANL, University of Oregon, and University of Maryland have developed the Cheetah and Savanna suite of workflow tools for running campaigns of workflows to explore online coupling of applications for reduction and analysis of large data. This work was performed as part of the Exascale Computing Project’s (ECP) CODAR (Co-design Center for Online Data Analysis and Reduction) project. This research introduces the Campaign abstraction for composing and executing workflows of workflows. It introduces a novel technique for binding application processes to compute resources. The toolset has been used for conducting research on different data reduction algorithms, and is being used for developing the EFFIS code coupling framework. The latest software release, Cheetah v.1.1.1, was released on June 30th, 2021.