In an AMD-financed research project two ph.d. students at Carolina State University in the USA published a paper where it has managed to boost processor performance up to 113% in AMD APU circuits, but iusing the integrated graphics circuit more efficiently.
The paper “CPU-Assisted GPGPU on Fused CPU-GPU Architectures” was published by students Yi Yang and Ping Xiang, and Mike Mantor from AMD, and shows a new technology that uses the graphics circuit inside the chip to boost overall performance. AMD is focusing on a more powerful graphics circuit with its latest x86 processors and recently revealed an even bigger focus on integrated circuits in the future. The big bottleneck of today is the support in software and technologies to use the extra computation power in regular everyday applications, so called GPGPU applications (General-Purpose computation on Graphics Processing Units).
The research was performed at Carolina State University and is an important step forward in this area and on average boost performance in AMD APU circuits with more than 20% by better usage of the integrated GPU cores.
“CPUs and GPUs fetch data from off-chip main memory at approximately the same speed, but GPUs can execute the functions that use that data more quickly. So, if a CPU determines what data a GPU will need in advance, and fetches it from off-chip main memory, that allows the GPU to focus on executing the functions themselves – and the overall process takes less time.”
The paper will be presened at the International Symposium on High Performance Computer Architecture, in New Orlean on February 27th in detail where the group reached up to 113% better performance, with on average 21.4% in tested applications.
The exact appllications it has teste, how this could be applied to everyay applications or for that matter if the technology would work with Intel’s integrated graphics is so far uncertain.
Source: Carolina State University
