In the article released today, research engineers at North Carolina State University have been able to squeeze out an average of 21.4 percent more performance through a new approach which improves fused processor performance.
This approach is specifically designed to optimize performance for AMD’s APU style of processors which have both the CPU cores as well as GPU cores on the same die. This research was funded by the National Science Foundation and AMD. As a result, there is little talk about how much performance could be improved on an architecture such as Intel’s which also has the CPU cores and GPU cores on the same die.
Associate professor, Dr. Huihang Zhou stated that, ?Chip manufacturers are now creating processors that have a ?fused architecture,? meaning that they include CPUs and GPUs on a single chip. 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.?
Interestingly enough, AMD and SCSU worked together to really harness the power of the GPU cores on the APU to improve the overall performance of the single chip. This is known as general purpose computing and really has been something that many people have been waiting for companies like AMD and Intel to be able to accomplish with their fusion architectures since there is die space sitting idle while the CPU is chugging along.
?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.This is more efficient because it allows CPUs and GPUs to do what they are good at. GPUs are good at performing computations. CPUs are good at making decisions and flexible data retrieval,? Zhou says.
We are extremely interested to see more details regarding this new way of processing data on a fusion architechture and this could potentially be the holy grail of computing for companies like AMD who are betting on fusion architectures. Hopefully these developments will be further supported by AMD with developer backing to help developers better make use of these improvements.
Here’s a link to the abstract of what work will be presented at the International Symposium on High Performance Computer Architecture in New Orleans on February 27th.