by Anton Shilov
01/19/2011 | 08:46 PM
Nvidia Corp. will integrate general-purpose ARM processing core(s) into a chip that belongs to Maxwell family of graphics processing units (GPUs), the company revealed in an interview. The Maxwell-generation chip will be the first commercial physical implementation of Nvidia's project Denver and will also be the company's first accelerated processing unit (APU).
"The Maxwell generation will be the first end-product using Project Denver. This is a far greater resource investment for us than just licensing a design," said Mike Rayfield, general manager of mobile solutions for Nvidia, in an interview with Hexus web-site.
Nvidia's initiative code-named Denver describes an Nvidia CPU running the ARM instruction set, which will be fully integrated on the same chip as the Nvidia GPU.
Nvidia Maxwell will be launched in 2013, it was revealed at Nvidia's GPU Technology Conference in September, 2010. Given the timeframe, it is logical to expect 20nm process technology to be used for manufacturing of Maxwell. The architecture due in almost three years from now will offer whopping 14 - 16GFLOPS of double-precision performance per watt, a massive improvement over current-generation hardware.
"Between now and Maxwell, we will introduce virtual memory, pre-emption, enhance the ability of GPU to autonomously process, so that it's non-blocking of the CPU, not waiting for the CPU, relies less on the transfer overheads that we see today. These will take GPU computing to the next level, along with a very large speed up in performance," said Jen-Hsun Huang, chief executive of Nvidia, at GTC 2010.
This is the first time when Nvidia publicly reveals timeframes for project Denver. Unfortunately, not all the details are clear at this point and it is unknown whether all members of the Maxwell family will have integrated GP ARM cores. General-purpose processing cores will bring mosts benefits for compute applications and therefore Nvidia may omit ARM from low-cost designs.