Researchers in the North Carolina State University have developed a method for increased interaction between a central processor (CPU) and graphics processor (GPU) as part of an accelerated processor (APU) to increase performance by over 20 per cent.
The new research has shown a method for handing off certain instructions to the different internal parts, depending on what sort of calculations are required. The CPU takes care of pre-fetching data while the GPU with its improved computation power handles the more complex calculations.
"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," said the co-author of the study, Dr Huiyang Zhou.
While both central and graphical processors can retrieve information from memory at about the same speed, GPUs perform calculations and a far more impressive rate, meaning it makes perfect sense to use a CPU as a baggage handler for its more powerful companion. With the CPU handling the data collection, it frees up extra computational cycles for the graphics chip.
The research paper with the results of the study is set to be released late February, at that point pehaps allowing AMD and Intel to begin applying the discovered techniques as part of their future developments. If the 20 per cent plus performance gains are anything close to accurate, it could open up a whole new jump in computing power with upcoming generations of hardware.