Knox Stories
Knox Day of Dialogue Continues Annual Tradition of Creating Meaningful Conversations
During his keynote, Wall asked attendees to reflect on why they believe everyone should be valued and respected.
Knox College has been awarded a quarter-million-dollar grant by the National Science Foundation for research into making super-computers more efficient. The work, directed by David Bunde, associate professor of computer science, focuses on "task mapping" -- assigning parts of a computer program to networked processors, boosting the efficiency of high performance computers.
The three-year grant of $258,631 includes stipends for a number of student assistants, engaging them in leading-edge research and preparing them for leadership positions in science, engineering and technology.
The grant will allow six students to work with Bunde at Knox through the summer. The grant also provides funding for two students to work on the research during the academic year, and one student during Knox's winter break. It also supports travel to present research findings at academic conferences.
The project, formally titled "Task Mapping for Emerging Network Topologies," aims to improve the performance of large scientific simulations, such as weather prediction and rocket design.
Bunde's prior work in the field has shown that task mapping can boost processing speed by up to 30%.
"Today's multi-processor computer systems are so large that we're running into power limitations," Bunde says. "New forms of networking are being developed, and task mapping has the potential to improve network performance. It arranges the communicating parts of the application in a way that distributes messages more evenly through the network, preventing any part of it from becoming overloaded. By improving application performance on next-generation computer systems, the project will help realize the full potential of these powerful systems."
Above, file photos of David Bunde and his students in earlier phases of his research into task mapping and other ways to improve the efficiency of high performance computer systems.
Published on September 22, 2014