Parallel Machines Commissioned|
Posted: 21 March 1999
On 19 March 1999, the Department of Computer Science commissioned the SGI Reality Monster and PixelFlow, two significant components of our parallel graphics computing facilities. Faculty, staff, students, and Candidates Day participants gathered in the Machine Room for the ceremony. Also attending from University administration were Risa Palm, dean of the College of Arts and Sciences, and Richard J. Richardson, provost, along with Jacqueline Resnick, director of Proposal Development Initiatives. Kevin Hidalgo, a locally-based sales representative from Silicon Graphics, Inc., also attended.
After opening remarks by Stephen F. Weiss, professor and chair, Henry Fuchs, Federico Gil professor, discussed the significance of the two machines. Our Reality Monster is one of only three such machines to be housed on a university campus, and is the only one in use as a laboratory machine. Henry noted that PixelFlow is unique: it is the only machine of its kind anywhere in the world. Henry praised the many current and former Department researchers who worked to make PixelFlow a reality. He thanked Department and University administrators for allowing such a favorable research climate to flourish.
Developed by Silicon Graphics, Inc., the Reality Monster, an Onyx2(TM) Infinite Reality2(TM) workstation, has 32 processors, 16 gigabytes of main memory, eight raster managers, and eight InfiniteReality2 graphics subsystems. It has been named Evans, in honor of computer graphics pioneer and educator, David C. Evans.
PixelFlow, a scalable parallel graphics engine, grew out of an idea by John G. Eyles, research associate professor, Steven Molnar (Ph.D. 1991) adjunct assistant professor, and John Poulton, research professor. It was designed and built by researchers in our Department, Division PLC, and Hewlett-Packard Corp.'s Chapel Hill Graphics Lab. PixelFlow machines can be as small as three nodes or much larger. We have assembled configurations as large as 36 nodes. Each node has two PA-8000 processors and 128 megabytes of main memory, as well as a 128 x 64 SIMD processor array and 64 megabytes of texture memory.
Claire L. Stone
Publications & Publicity Manager
|Department of Computer Science
Campus Box 3175, Sitterson Hall
College of Arts & Sciences
The University of North Carolina at Chapel Hill
Chapel Hill, NC 27599-3175 USA
Phone: (919) 962-1700
Fax: (919) 962-1799