NCSU researchers get on-chip GPUs and CPUs to cooperate

Manufacturers are moving to consolidate more and more functions on a single chip. As Dr. Elan Spillinger, vice president for hardware and technology at the Microsoft Interactive Business Entertainment Unit, said at a DesignCon keynote address January 31, designs used to incorporate a CPU, GPU, Northbridge, and Southbridge, but now that can all be incorporated in a single SoC. Microsoft takes credit for being the first company to combine a CPU and GPU on one chip.

But manufacturers aren't getting all the performance they could out of such combination chips, according to  researchers from North Carolina State University, who say they have developed a new technique that allows graphics processing units (GPUs) and central processing units (CPUs) on a single chip to collaborate – boosting processor performance by an average of more than 20%.

“Chip manufacturers are now creating processors that have a ‘fused architecture,’ meaning that they include CPUs and GPUs on a single chip,” says Dr. Huiyang Zhou, an associate professor of electrical and computer engineering. “This approach decreases manufacturing costs and makes computers more energy efficient. However, the CPU cores and GPU cores still work almost exclusively on separate functions. They rarely collaborate to execute any given program, so they aren’t as efficient as they could be. That’s the issue we’re trying to resolve.”

He adds, “Our approach is to allow the GPU cores to execute computational functions, and have CPU cores pre-fetch the data the GPUs will need from off-chip main memory.”

The approach will be described in the paper to be presented February 27 at the 18th International Symposium on High Performance Computer Architecture, in New Orleans. The paper was co-authored by NC State Ph.D. students Yi Yang and Ping Xiang and by Mike Mantor of AMD. The research was funded by the National Science Foundation and AMD.