Carole-Jean Wu’s research over the last few years has focused on computer chip designs for laptop computers, desktops and servers. As computers move to more cores—dual-core, quad-core—the challenge is how to make the different cores communicate with each other as efficiently as possible.
“Computer users open different applications—listening to music, doing homework on MATLAB, watching a video—and expect them to run very smoothly. Without careful management, these programs can fight for resources such as memory space,” she says.
One of Wu’s recent research projects looked a critical piece of this process: cache memory.
“You can think of it as a common storage space that different applications all use,” Wu says.
Wu’s research observed the amount of resources that various applications require then fed this information into a predictor that is sitting next to the cache memory. By predicting which applications could accelerate performance with a greater allocation of cache resources, overall performance can be improved.
Since joining Arizona State University as assistant professor in the School of Computing, Informatics, and Decision Systems Engineering in fall 2012, Wu has expanded her research to the mobile field.
Mobile phone chip makers are putting more cores on the same chip to support an increasing number of capabilities far beyond just making calls. In addition to the multi-core challenges faced on computing devices, mobile devices have an added complication with more specialized hardware.
“While we are putting general purpose cores capable of handling a wide range of applications on mobile phones, there are also specialized units such as the 3G antenna,” Wu explains.
Her goal is to devise a communication model such that these heterogeneous devices will be able to talk to each other more efficiently—not only for performance, but to improve the user experience and gain greater energy efficiency.
Another area of interest is how to create efficiencies for new chips that include a graphics engine. Since the graphics engine is sitting next to the CPU, she says many researchers are studying how to offload applications and exploit the additional hardware resources. For mobile users, this could mean the ability to gain a more rich, high-definition video experience, for example.
Wu says that growing up, she was always interested in building things.
“A lot of what we build in this field is inside the computer, so it may not seem like we are building something with our hands. But I can see it being compiled, and see things running and moving,” she says.
Wu interned with Intel, IBM and Google. She was the recipient of the 2011-2012 Intel Ph.D. Fellowship Award. She says the experience helped her to understand what industry is thinking and how that relates to knowledge gained through study and research.
Ultimately, she pursued a doctoral degree because she says she really enjoys teaching. She earned her Ph.D. and M.A. degrees in electrical engineering from Princeton University and holds a B.S. in electrical and computer engineering from Cornell University.
For more on Wu’s research, publication and other activities, visit: faculty.engineering.asu.edu/carolewu.
