CS Seminar: An Ecosystem for the New HPC: Heterogeneous Parallel Computing by Dr. Wu Feng
With processor core counts doubling every 18-24 months and penetrating all markets from high-end servers in supercomputers to desktops and laptops down to even mobile phones, we sit at the dawn of a world of ubiquitous parallelism, one where extracting performance via parallelism is paramount. That is, the "free lunch" to better performance, where programmers could rely on substantial increases in single-threaded performance to improve software, is over. The burden falls on developers to exploit parallel hardware for performance gains.
Osman Balci receives SCS service award
Publish Date: 09/12/2013
Osman Balci has been awarded the McLeod Founder's Award for Distinguished Service to the Profession by The Society for Modeling and Simulation International (SCS). SCS states "the awards in this category are intended to recognize outstanding contributions to the science and technology of the modeling and simulation process." SCS is the world’s premier professional society devoted to modeling and simulation.
Seminar Series - Mash Sosonkina: Achieving Energy Savings in Message-Passing Applications
Mash Sosonkina
Friday September 13
Torgersen 2150
11:15am-12:30pm
Distinguished Lecture - Innovating for Society: Realizing the Transformative Impact of Computing and Communication
Dr. Farnam Jahanian
Location: 2150 Torgersen Hall
Date: Friday, September 20, 2013
Time: 11:15-12:30pm
This talk is open to the general public.
User-Intention Based Android Malware Detection, poster
Trust-based Service Composition and Binding for Tactical Networks with Multiple Objectives
A Partition Function Approximation Using Elementary Symmetric Functions
VT GPU Computing Workshop a sold out success
Publish Date: 09/02/2013
The Virginia Tech Department of Computer Science and NVIDIA collaborated to host a hands-on, two-day workshop to survey the broad range of GPU-accelerated applications across all domains of scientific and engineering research. The August 14-15 workshop attracted more than 120 attendees from four institutions and 18 different departments. Participants included graduate students, postdocs, researchers, and professors all learning firsthand how to program graphics processing units (GPUs) via the use of libraries, OpenACC compiler directives, and CUDA programming. Participants performed hands-on exercises to acquire the skills to use and develop GPU-aware applications.
Dr. Wu Feng of the Department of Computer Science began the workshop with an overview addressing heterogeneous parallel computing and an introduction to HPC environment and tools at Virginia Tech. Other presentations were made by Bob Crovella, who leads a technical team at NVIDIA and is responsible for supporting GPU computing products in the high performance computing ecosystem. Crovella gave an introduction to GPU computing and then dove into two more intense sessions, all hands-on by the participants.
Feng and Crovella addressed why heterogeneous computing and GPUs have become so important in sustaining and advancing the state of the art in scientific and research computing. Virginia Tech has its own GPU-accelerated supercomputer, HokieSpeed, designed and deployed by a team led by Dr. Feng, which debuted as the most energy-efficient commodity super computer in the U.S. in November 2011.