Where: Bldg 9. Lecture Hall I# 2322
Description
How can you change the world with access to inexpensive TF/s performance and PF/s leadership class capability? Are you interested in getting orders-of-magnitude performance increases over standard multi-core processors? Would you like that capability to scale across many devices as well? Multi-device code shows how to transition from a "Hello World" example to the largest leadership-class supercomputers. Start with a high-level software platform and delve down to lower APIs for performance or use of special features. See how a few lines of OpenACC source code can succinctly express parallelism yet compile to efficiently run on Fermi, Kepler, x86, plus AMD and Xeon Phi devices. Less is more! Utilize Thrust, CUDA, OpenCL, OpenMP and MPI for additional capabilities and performance. Working code demonstrates that GPUs can "see", read aloud, plus interactively compute and visualize 100 million+ 3D LiDAR data points. http://insidehpc.com/2011/11/22/farber-book-on-cuda-serves-up-easy-teraflops/ Simplifying Portable Killer Apps with OpenACC and CUDA-5 Concisely and Efficiently Clicking GPUs into a Portable, Persistent and Scalable Massive Data Framework Workshop attendees need to be able to login, edit, compile and run C/C++/CUDA programs.
Michael Young
Michael Young is an American geneticist who contributed to the discovery of molecular mechanisms that regulate circadian rhythm, the 24-hour period of biological activity in humans, and other organisms. He was awarded the 2017 Nobel Prize in Physiology or Medicine. Young received his PhD in genetics in 1975 from UT Austin. Following a postdoc in biochemistry at Stanford University School of Medicine, he was appointed assistant professor at Rockefeller University. He was named professor and later, he was appointed the university’s vice president for academic affairs and Richard & Jeanne Fisher Professor.
Rob Farber
Rob Farber is a recognized expert in highly parallel programming with emphasis that code is portable, scalable and independent of architecture. He is currently Chief Scientist for his consulting company that has a variety of research grants plus contracts with large technology companies (including NVIDIA, Intel and AMD). He has been a scientist at US national laboratories including Los Alamos National Laboratory, Lawrence Berkeley National Laboratory, and Pacific Northwest National Laboratory and on-staff as a visiting scientist at international institutions such as the Irish Center for High end Computation. As an entrepreneur, he has co-founded two successful companies and performed as consultant to fortune 100 companies. Rob also served on the external faculty at the Santa Fe Institute and authored “CUDA Application Design and Development”, with the follow-up book in development. As an author/teacher, Rob has 100+ peer reviewed scientific publications. His freely available teaching codes have ran on two of the largest computer systems currently available: the Xeon Phi based Stampede system at UT Austin (2.2 PF/s using 3,000 nodes), and the Nvidia Kepler based Titan system at Oak Ridge National Laboratory(13 PF/s using 16,384 nodes).
No resources found.
No links found.
