The ECX Project
Optimizing the Energy Usage and Cognitive Value of Extreme Scale Data Approaches
Scientific discovery at the extreme scale is a unique technical challenge, requiring the reduction of massive amounts of data into compact analysis products that capture key scientific insights. This analysis process needs to occur under extreme scale computational constraints including minimizing 1) data movement, 2) energy usage and 3) storage usage. Put simply, extreme scale computing platforms are to achieve a three orders of magnitude increase in computational performance while consuming only two times the electrical power of current platforms. Data movement costs will dominate energy usage at this scale, so the HPC community expects extreme scale analysis algorithms will be utilized to reduce simulation results insitu – that is, during the simulation run. This reduction will occur, broadly speaking, via some type of adaptive sampling, such as signal, statistical or feature-based sampling.
We are investigating how changes in our sampling algorithms – necessary because of exascale power constraints – impact the cognitive value of the resulting data. Our goal is perceptual and cognitive optimization of tools that enhance analytic workflows while minimizing power consumption. We will pursue a three phase approach to the research, isolating specific portions of the potentially vast experimental space, in order to be able to draw conclusions across scales – from supercomputers to compute nodes, down to the subsystems of the compute node. We have assembled a team of experts from the diverse areas of Algorithms, Power, and Cognition/Perception to address the issues inherent in the optimization of the human/compute system at exascale.
This website is an ongoing document of the people, the research and the results from the ECX project.
The ECX project has resulted in several online communities that include a variety of open source software, websites, interactive tools, and information that you can take advantage of. We invite you to use these tools, and add to them as well.
- The Evaluation Toolkit at ETKlab (www.etklab.org)
- An open source toolkit designed to help you conduct image-based perceptual studies on your own visualizations.
- The ECX Experimental Test Harness (ETH) (Experimental Test Harness GitHub)
- A lightweight toolkit for exploring data analysis and visualization workflows across a large parallel design space.
- sciviscolor (sciviscolor.org)
- This website collects a range of research on colormaps, color schemes, and interactive creation of tuned colormaps.
- colormeasures.org (colormeasures.org)
- discoveryjam (www.discoveryjam.com)
- A hack-a-thon inspired way for scientists, designers and visualization experts to collaborate on solutions for complex data analysis problems.
The ECX project has a GitHub repository. All of the open-source code released under the ECX grant can be found at https://github.com/ascr-ecx
Please email email@example.com with questions about this work, and any of the related websites.