We’re in Seattle this week attending the US DOE Workshop on Modeling and Simulation of Exascale Systems and Applications — “ModSim’13” [www]. A copy of Rich’s talk and pointers to supplemental materials is available here: hpcgarage.org/modsim13
Rich is back from the University of Vienna, where he had the great fortune to serve as an external referee on Hana Strakova’s PhD dissertation committee. Hana was a Fall 2011 intern in our lab; her primary thesis advisor is Wilfried Gansterer. Her thesis breaks new ground in the development of resilient and asynchronous numerical linear algebra algorithms for highly unreliable systems, using so-called “gossiping” techniques: “Truly distributed approaches to orthogonalization and orthogonal iteration on the basis of gossip algorithms.” Naturally, she passed with flying colors. Congratulations!
Left-to-right: Rich, Dr. Hana Strakova, and her advisor, Wilfried Gansterer (University of Vienna).
We are in Warsaw this week for the International Conference on Parallel Processing and Applied Mathematics (PPAM) [www]. If you are attending, be sure to see Marat’s talk and Rich’s talk:
- Marat Dukhan and Richard Vuduc. “Methods for high-throughput computation of elementary functions.” Tue., Sept. 10, Track A, 10:20am – 12:25pm. [Marat’s slides: PDF (5.3 MiB)]
- Richard Vuduc. “How much (execution) time, energy, and power will my algorithm cost?” Wed., Sept. 11, Invited Keynotes, noon – 1:20pm. [Rich’s materials: Click here]
Beyond these, there are two additional not-to-be-missed Georgia Tech talks: Anita Zakrzewska, a student of David Bader, and Jeff Vetter (ORNL+GT):
- Jeff Vetter. “Exploring Emerging Technologies in the HPC Co-Design Space.” Tue., Sept. 10, Invited keynote track A, 2:00-3:20pm.
- Anita Zakrzewska and David Bader. “Measuring the sensitivity of graph metrics to missing data.” Wed., Sept. 11, Track A, 10:20am – noon.
Copernicus Science Center in Warsaw, where we held the first day of PPAM’13.
Marat is at the BLIS Retreat, and just gave a presentation on Peach-Py, his Python-based assembler. Peach-Py gives you the control and efficiency of assembly programming with the convenience and ease of Python syntax, while also automating key tasks like constant and register allocation. Peach-Py currently supports a variety of x86 and ARM instruction sets.
If you missed his talk, you can see the PDF slides (and eventually the video) here: PDF [357 KiB]
Incidentally, BLIS is an exciting new effort to build a new infrastructure for high-performance primitives.
Peach-Py is a Python-based assembler. It improves assembly-level programming by giving the programmer much of the control and power of assembly combined with the convenience and ease of Python syntax, all while automating a number of key tasks, like register allocation.
Marat is at Hot Chips 2013 this week presenting his poster entitled, “What a fast FPU means for algorithms: A story of vector elementary functions.” This work lays the foundations for Yeppp!, his new numerical library that embraces the “flops-are-free” mantra to develop new high-speed implementations of functions like log, exp, sin, cos, among others, and polynomial evaluation. A copy of his poster is available here [PDF, 497 KiB].
Marat’s Hot Chips 2013 poster (thumbnail)
Rich is at the American Geophysical Union this week for a US Department of Energy workshop on “Exascale Mathematics.” You can read his position paper and talk materials [here].
Sangmin is in Lugano, Switzerland this week to present his new paper on Griffin at ISSTA, the 2013 International Symposium on Software Testing and Analysis. Griffin is a tool that aims to help programmers find bugs in their concurrent software. Congrats, Sangmin!
- Sangmin Park, Mary Jean Harrold, and Richard Vuduc. “Griffin: Grouping suspicious memory-access patterns to improve understanding of concurrency bugs.” In Proceedings of the International Symposium on Software Testing and Analysis (ISSTA), Lugano, Switzerland, July 15-20, 2013.
A bug graph, one of Griffin’s aids for programmers who need to debug their concurrent software.
Congratulations to David S. Noble, Jr., our most recent alumnus (CSE MS Spring 2013), who just landed a job at CD-adapco, makers of scalable high-quality software for computational fluid dynamics applications. To celebrate, Dave unveils his highly graceful swan-like dunking ability in this 3-second video clip (~ 6 MiB QuickTime movie): Dave_Dunk_Attempt.
Note: This skill is a requirement to graduate from The HPC Garage, which is how you can tell Kent is not yet ready (~ 15 MiB .mp4, 8 seconds): Kent_Dunk_Attempt
Jee and Kent are speaking on Wednesday, May 22 at IPDPS’13. If you are attending, be sure to check out their talks!
- Jee Choi et al. “A roofline model of energy.” Session 14, 10am-noon.
- Kenneth Czechowski and R. Vuduc. “A theoretical framework for algorithm-architecture co-design.” Session 17, 1:30-3:30pm.
Both sessions are in the William Dawes Room at the Hyatt Regency Cambridge. Hope to see you there!
Rich gave a keynote at the 3rd International Workshop on Accelerators and Hybrid Exascale Systems (AsHES) earlier this week. A copy of his slides is available here: hpcgarage.org/ashes2013
A notional space of processors.
Kent and Jee will each present a paper at IPDPS’13, to be held in Boston in May. Kent’s paper is about a unique take on high-level algorithm-architecture co-design; Jee’s paper is about his “roofline” energy model. If you plan to attend, be sure to see their talks! We’ve also posted preprints under Papers.
- Jee Choi and Richard Vuduc. A roofline model of energy. In Proc. IEEE Int’l. Parallel and Distributed Processing Symp. (IPDPS), Boston, MA, USA, May 2013. [PDF | BibTeX]
- Kenneth Czechowski and Richard Vuduc. A theoretical framework for algorithm-architecture co-design. In Proc. IEEE Int’l. Parallel and Distributed Processing Symp. (IPDPS), Boston, MA, USA, May 2013. [PDF | BibTeX]