UpDate - Vol. 13, No. 35, Page 12
June 23, 1994
Professor's research attracts national attention

     Holding up a computer backplane, Phillip Christie, electrical
engineering, shows a maze of tiny wires which he compares to
spaghetti. The other side of the board holds silicon chips.
     As more devices, such as chips and transistors are packed into
computers, the number of wires that connect these devices multiplies
and, ultimately, becomes the limiting factor in performance.
     The importance of these connecting wires has not been fully
appreciated, especially in cases where there is a need to scale up the
size of an existing computer design.
     Beyond a certain point, scaling up is impossible to do, Christie
said.  His research, on the wiring component of computers, was
recently written up in the Electronic Engineering Times.
     Christie is approaching this problem from a different
perspective. "It's like gas molecules enclosed in a box," he
explained.
     "With three or four you can see how they react, bouncing off each
other or the sides of the enclosure. As you add more molecules, this
becomes increasingly impossible, so you use other techniques to
predict what is happening, such as plotting temperature and pressure
and looking at the whole," he said.
     Christie views computer wiring as a collective whole and is
applying fractal (fractional dimension) theory to the problem.
Fractals are endlessly self-repeating shapes that are mathematical
expressions of the real world, which is not just points, lines and
curves but irregular shapes such as clouds and coast lines. Fractals
can have in-between or additional dimensions.
     He is also making use of a concept called "renormalization group
theory" and other theories to develop a new mathematical approach to
computer system modeling, particularly the wiring component.
     The "renormalization group theory," for which physicist Ken
Wilson received the Nobel Prize, has been used to predict the behavior
of complex systems, such as systems of subatomic particles. Christie
is applying this theory to the development of a new approach to
computer system modeling.
     Electrical engineering is not static but full of energy.  It's a
rapidly changing field, according to Christie.
     For example, he said, teaching a graduate course on the theory of
semiconductor devices is interesting because the course content can
change in six months' time.
     "In the past few years, there have been three Nobel Prize winners
in the field, and our understanding increases and changes with new
research. To keep current requires that engineers be involved in
research themselves," he pointed out.
     "The department, as a whole, is doing cutting-edge research in
computers which could have considerable impact in the electrical
engineering field," Christie said.
                                                   -Sue Swyers Moncure