CUDA GPGPU Parallel Computing Newsletter - Issue 64

GPU-Accelerated Real Science
This week’s Spotlight is on Dr. Jeffrey Vetter of Oak Ridge National Laboratory and Georgia Tech. Here’s an extract from our interview.

NVIDIA: Jeff, what is the focus of your work?

Jeff: Over the last decade, I have been investigating hardware and software technologies that will most likely appear in future supercomputer systems, and which of those technologies best satisfy specific application workloads. I have worked on a number of projects: IBM BlueGene/L, Cray X1, Cray XT, FPGAs, GPUs and other technologies. Our team’s early work has contributed to the design and deployment of the NSF Keeneland system and the DOE Titan system.

Not surprisingly, our work over the last several years has primarily focused on GPUs. Our team is involved in most every aspect of GPUs in computational science: future architectures, programming systems, applications development, and education and outreach. For example, we are partners on NVIDIA’s Echelon research project, which is sponsored by the DARPA UHPC program with the goal of fitting one PetaFLOPS in one rack and using less than 57K watts of power….

NVIDIA: What are the key drivers in supercomputing today?

Jeff: From the facility or data center perspective, the key driver is the energy required for running large scale supercomputers. This is not just a prediction. Look at most contemporary supercomputing facilities: they are often limited by the amount of power that can be physically delivered to the computer in the building….

On the other hand, applications developers are most concerned about programmability. The last significant transition for the scientific computing community was the transition in the 1990s from vector computing to distributed memory computing with MPI. In order to provide solutions to these questions, our team is investigating multiple fronts: CUDA, compiler directives, runtime libraries, frameworks and debugging and correctness tools. It is an exciting time to be in computer science!

NVIDIA: How does CUDA fit into the modern computing landscape?

Jeff: CUDA is a phenomenon. In less than five years, the CUDA programming model has grown from its initial introduction to wide adoption. It is easy to forget how challenging it was to program GPUs prior to CUDA. These days, CUDA is so pervasive that many students get their first introduction to parallel programming and fine-grained parallelism with CUDA on their laptop GPU….

- Read the complete interview with Jeff Vetter

(To suggest a CUDA Spotlight, email cuda_week_in_review@nvidia.com)

NVIDIA GPU Technology Theater at SC11

back to the top

Going to SC11 in Seattle? If so, drop by the GPU Technology Theater at the NVIDIA booth (#2719). The Theater will feature talks from industry experts including Jack Dongarra, University of Tennessee; Patrick McCormick, Los Alamos National Laboratory; Richard Brower, Boston University; and Satoshi Matsuoka, Tokyo Institute of Technology. For the complete speaker lineup, visit the NVIDIA SC11 event page. The talks will be streamed live via Facebook at http://apps.facebook.com/nv-supercomputing/.

Oak Ridge National Lab to Deploy Titan
Oak Ridge National Laboratory (ORNL), which operates the world's premier open science computing facility for the U.S. Department of Energy, will deploy a new supercomputer called Titan, based on Tesla GPUs. The system, which is referenced in today's CUDA Spotlight with Dr. Jeff Vetter, will be used across fields such as materials science, energy technology, medical research and more.
- See the VentureBeat interview with NVIDIA's Steve Scott (8 mins.): http://youtu.be/IkC8YRo9Mts

GTC U.S. Submission Deadline
If you (or someone you know) is interested in speaking at GTC U.S. (May 2012), proposals must be submitted by Thurs., Nov. 3, 2011. Details can be found on the GTC website.
- To make a submission, visit: https://gtc-submissions.confreg.com/

Cornell Launches Red Cloud
The Cornell University Center for Advanced Computing (CAC) has launched Red Cloud, an on-demand research supercomputing service available by subscription. The basic offering, called "Red Cloud," is an Infrastructure as a Service (IaaS) that runs Eucalyptus, the open source cloud computing platform. The second offering, "Red Cloud with MATLAB," is a Software as a Service (SaaS) that runs MATLAB Distributed Computing Server and features NVIDIA GPUs. Red Cloud services run on Dell PowerEdge C servers.
- Visit www.cac.cornell.edu/redcloud

Congratulations to Francesco Rossi
Congratulations to Francesco Rossi, University of Bologna, on his recent master’s thesis titled: "Development of Algorithms for an Electromagnetic Particle in Cell Code…." While completing his thesis, Francesco developed Jasmine, a flexible CPU+GPU PIC (particle in cell) framework to assist in the design of laser-plasma accelerators. Jasmine runs PIC 3D simulations on multi-GPU clusters, such as CINECA’s Tesla-powered PLX, gaining large speedups.
- Read more: http://www.roxlabs.com/?page_id=110

Intergraph: Video for Geo-Spatial Awareness
At the recent GeoINT trade show, Intergraph announced a new release of its GeoMedia Motion Video Analyst solution with CUDA acceleration. GeoMedia Motion Video Analyst speeds the process of creating images from full motion video (FMV). This feature improves the ability of defense analysts to maintain situational awareness and act quickly on intelligence data.
- For info, see: http://bit.ly/pBjGMX

back to the top

Webinar: Speed Up MATLAB With NVIDIA GPUs, by Dan Doherty, MathWorks
NVIDIA Engineering Strike Force, by Brian Kelleher
NVIDIA CEO Wraps Up All Things D Asia Event, by Bob Sherbin

back to the top

NEW: Each week we highlight a session from a GPU Technology Conference event. Here is our pick for this week: Computer Vision on GPU with OpenCV (Israel 2011) by Anton Obukhov, NVIDIA
- For info, see: http://bit.ly/r5Qay6

back to the top

NEW: GE Healthcare is seeking a software engineer to specify, design and implement software for a new X-ray acquisition system. Desired characteristics include experience with GPU/CUDA and imaging software.
- See: http://jobs.gecareers.com/job/Salt-Lake-City-Software-Engineer-II-HC-Job-UT-84101/1284204/
(To submit a job listing, email cuda_week_in_review@nvidia.com)

back to the top

The GPU Meetups offer a great way to learn about GPU computing and meet interesting people in a relaxed environment. Please feel free to attend any of these upcoming Meetups:

GPU Meetup of Boston, Thurs., Nov. 3 at 6:00 pm

Topic: HPC and Protein Simulation, Francesco Pontiggia, Brandeis University

GPU Meetup of Silicon Valley, Mon., Nov. 7 at 6:15 pm

Topic: GPU Programming with Thrust and Copperhead, Bryan Catanzaro, NVIDIA

GPU Meetup of Seattle, Wed., Nov. 16, 7:00 pm (networking), 7:45 pm (program)

Special SC11 Meetup! Talks by NVIDIA, Microsoft, LexisNexis. Location: Amazon

GPU Meetup of New Mexico, Wed., Nov. 16, 7:00 pm

Topic: ROMIO and MPI-IO in Hybrid HPC

GPU Meetup of Brisbane, Nov. 17, 6:00 p.m

GPU Meetup of New York City, Nov. 21, 6:00 pm

(Would you like to start a Meetup? Email cuda_week_in_review@nvidia.com)

back to the top

GTC Asia 2011 (Dec. 14-15)
Poster deadline: Nov. 3
http://www.gputechconf.cn/en/call-for-posters.html

GTC U.S. 2012 (May 14-17)
Session deadline: Nov. 3
Poster deadline: Dec. 8
http://www.gputechconf.com/page/participate.html
https://gtc-submissions.confreg.com/

November 2011	back to the top

- NEW: Accelerate Science to Treatment with the MD SimCluster (Webinar)

Nov. 8, 2011

Sponsored by NOVATTE, A*Star and NVIDIA

For info, email to Oxana Plis, op (at) novatte.com

- Supercomputing 2011 (SC11)

Nov. 12-18, Seattle, Washington

Learn about NVIDIA activities at SC11: http://www.nvidia.com/sc11.html

For more information on SC11, visit http://sc11.supercomputing.org/

Join the GPU Technology Theater from your desk via the SC11 Facebook live stream: http://apps.facebook.com/nv-supercomputing/

- GPU Programming for Defense/Intelligence — AccelerEyes (Webinar)

Nov. 15, 2011

Learn to accelerate common defense and intelligence algorithms using easy, powerful programming libraries, with Jacket for use with MATLAB and LibJacket for C/C++/Fortran.

http://bit.ly/rdZ8pH

- NEW: Heterogeneous Data-Parallel Programming (Webinar)

Nov. 16, 2011

Presenter: Prof. Satnam Singh, University of Birmingham, U.K.

https://www2.gotomeeting.com/register/164396834

- CUDA Training (Basic and Advanced) — CAPS

Nov. 22-24, 2011, Rennes, France

http://t.co/DWe0Zqka

Email: training (at) caps-entreprise.com

- NEW: CUDA 4-Day Training Course – Acceleware

Nov. 22-25, 2011, Frankfurt, Germany

Presented by Acceleware with Microsoft

Instructor: Michael Durocher

http://www.acceleware.com//nov22munich

December 2011

- AGU (American Geophysical Union) Meeting

Dec. 5, 2011, San Francisco

Session on High-Res Modeling Using GPU and Many-Core Architectures

http://sites.agu.org/fallmeeting/scientific-program/session-search/530

- NEW: Intro to GPU Programming Workshop - La Maison de la Simulation

Dec. 5-9, 2011, France

http://www.maisondelasimulation.fr/index.php

- GTC Asia

Dec. 14-15, 2011, Beijing, China

Featuring the latest GPU computing breakthroughs, demos and presentations.

http://www.gputechconf.cn/home.html

- LibJacket CUDA Library for Maximus — AccelerEyes (Webinar)

Dec. 15, 2011

Learn to integrate computations with visualizations in a CUDA-based app through simple visualization functions for plotting, image and volume rendering, and more.

http://bit.ly/rdZ8pHs

(To list an event, email: cuda_week_in_review@nvidia.com)

Tesla MD SimCluster	back to the top
– Want to test drive a GPU? Try the Tesla Molecular Dynamics SimCluster: www.nvidia.com/simcluster
Downloads
– CUDA 4.0: www.nvidia.com/getcuda – Parallel Nsight: http://developer.nvidia.com/nvidia-parallel-nsight
Webinars
– CUDA: www.nvidia.com/webinars – Parallel Nsight: http://developer.nvidia.com/developer-webinars

CUDA Registered Developer Program

– Sign up: www.nvidia.com/paralleldeveloper

CUDA GPUs

– List of CUDA-enabled GPUs: http://developer.nvidia.com/cuda-gpus

CUDA on the Web

– See previous issues of CUDA: Week in Review: http://is.gd/cBXbg
– Follow CUDA & GPU Computing on Twitter: www.twitter.com/gpucomputing
– Network with other developers: www.gpucomputing.net
– Stayed tuned to GPGPU news and events: www.gpgpu.org
– Learn more about CUDA on CUDA Zone: www.nvidia.com/cuda
– Check out the NVIDIA Research page: www.nvidia.com/research

CUDA Recommended Reading

– Future of Computing Performance: http://bit.ly/hYqH2H
– Supercomputing for the Masses, Part 21: http://is.gd/Fj56gf
– CUDA books: http://www.nvidia.com/object/cuda_books.html

CUDA Recommended Viewing

– The Third Pillar of Science: www.nvidia.com/object/race-for-better-science.html

– GTC 2010 presentations: www.nvidia.com/gtc

– SC10 presentations: www.nvidia.com/object/sc10_theater.html

CUDA is a parallel computing platform and programming model invented by NVIDIA. It enables dramatic increases in computing performance by harnessing the power of the graphics processing unit (GPU). NVIDIA provides a complete toolkit for programming on the CUDA architecture, supporting standard computing languages such as C, C++ and Fortran as well as APIs such as OpenCL and DirectCompute. Send comments and suggestions to: cuda_week_in_review@nvidia.com

Click here to opt in specifically to CUDA: Week in Review.