The SciDAC RAPIDS Institute, along with our sibling Institute FASTMath, has been renewed for a new five-year term under SciDAC-5. RAPIDS and FASTMath continue to serve important roles in connecting science teams with computer scientists and applied mathematicians and developing new technologies to aid in their work.
Both Institutes have incorporated additional AI capabilities into their portfolios in recognition of the importance of this growing area, while RAPIDS retains capabilities to assist in adapting to new platforms, managing scientific data, and visualizing and analyzing large volumes of scientific data.
I’m incredibly excited to continue to lead this talented team with my co-lead Lenny Oliker.
If you have been tracking the activities of our team, you’re aware that we’ve been working in the area of distributed data services for some time, along with colleagues at Los Alamos, Carnegie Mellon, and the HDF Group. Under the name Mochi, we’ve been working towards a methodology and set of components for constructing these services in a way that is productive to develop, very flexible, and high performing on modern hardware.
I’m very excited that an overview of this activity was recently published in the Journal of Computer Science and Technology, and I hope the paper will help others understand both the scope of challenges and our contributions in this area.
As I mentioned in a prior post, back in September 2018 I had the opportunity to lead a second workshop to discuss storage systems and I/O as relates to the DOE mission. ASCR has posted the report on the OSTI web site. I really appreciate that so many of my colleagues were willing and able to spend their time and energy participating in this activity, and I think the outcome is very informative.
Many thanks to Lucy Nowell for sponsoring the meeting and to Laura Biven for helping shepherd the report to completion.
Back in September I had the opportunity to lead a second workshop to discuss storage systems and I/O as relates to the DOE mission. Recently I presented a summary of the findings from this workshop to ASCAC, who advises DOE on topics related to ASCR.
With increasing heterogeneity in systems and in the workflows of DOE science and mission teams, there are many opportunities for CS research to contribute to more effective solutions than are available today. I’m hopeful that my team, my collaborators, and I will be able to continue our work in this important area.
At the Scientific Discovery through Advanced Computing Principal Investigator (SciDAC PI) Meeting I presented an overview of the RAPIDS Institute. This presentation summarizes the capabilities in the Institute as well as serving as a “map” to points of contact for the technologies we can provide to DOE scientists.
This is the first PI meeting since we were established, so it’s an exciting chance for us to meet with the science teams and learn about their challenges and how we might assist them.
Barb Helland, Associate Director of Advanced Scientific Computing Research, announced the formation of the RAPIDS Institute as part of her presentation to ASCAC.
The objective of RAPIDS is to assist Office of Science (SC) application teams in overcoming computer science and data challenges in the use of DOE supercomputing resources to achieve science breakthroughs. To accomplish this objective, the Institute will solve computer science and data technical challenges for SciDAC and SC science teams, work directly with SC scientists and DOE facilities to adopt and support our technologies, and coordinate with other DOE computer science and applied mathematics activities to maximize impact on SC science. I’m serving as Director of RAPIDS, with Lenny Oliker serving as Deputy Director and many talented researchers contributing their skills. See the web site for more information.
The Greater Chicago Area Systems Research Workshop (GCASR) grew to 350 attendees from university and industry last year, bringing together the top systems researchers and industry technologists in the Chicago area. This year promises to be bigger and better!
Where: University of Chicago’s Logan Center
When: Wednesday, April 27, 2016
20 Expert speakers also present on leading technology and research on data science, deep learning and graphics hardware, cloud and data centers, security and reliability, scalable computing systems, and more!
- William J. Dally (Chief Scientist and SVP of Research at NVIDIA)
- Michael J. Franklin (Berkeley AMPLab Director and Professor, future CS Chair, UChicago)
- 10 Top Academic research presentations from University of Chicago, Northwestern University, Illinois Institute of Technology, Purdue University, University of Illinois, University of Michigan, and University of Wisconsin
- 9 Top Industry expert speakers from Argonne, EMC, Fermilab, Google, Groupon, HP, Huawei, Intel, and KCG
- 70 student research posters about leading research projects at Chicago/Midwest universities
Thanks to our sponsors: KCG, EMC, Intel, Chicago Innovation Exchange, and UChicago CERES Center.
ASCR just recently posted a call for proposals in the area of Storage Systems and I/O, following the release of the workshop report. There are two versions of the call, one for Labs and one for Universities and others.
We’ve started doing a series of highlight videos on research in our division. Many thanks to Jim Corsolini and Dan Harris for editing my rambling into a coherent story.
I don’t read a log of blogs. Adrian Coyler’s the morning paper is excellent. He summarizes a CS paper basically every week day, makes good (IMO) picks, and his summaries are very readable.
I doubt me linking to his blog is going to generate significant new traffic for him, but it’s the least I can do.