Nvidia Corp. said at a recent event that it would start shipping its next-generation graphics processing units (GPU) code-named Kepler by the end of the year. The company did not say that the new chips will actually become available, but it is highly likely that they will at least be formally introduced and the company starts to ship the new chips for revenue.
"Between the Fermi generation and Kepler, which we should start shipping by the end of the year, there is about 3x improvement in [double precision] performance per watt. [...] We are about to introduce [our next-generation GPUs]," said Chris Malachowsky, senior vice president of research and co-founder of Nvidia, at the company's GTC Workshop Japan event.
Kepler is Nvidia's next-generation graphics processor architecture that is projected to bring considerable performance improvements and will likely make the GPU more flexible in terms of programmability, which will speed up development of applications that take advantage of GPGPU (general purpose processing on GPU) technologies. Some of the technologies that Nvidia promised to introduce in Kepler and Maxwell (the architecture that will succeed Kepler) include virtual memory space (which will allow CPUs and GPUs to use the "unified" virtual memory), pre-emption, enhance the ability of GPU to autonomously process the data without the help of CPU and so on.
The new chip is projected to be made using 28nm process technology, even though Nvidia has not publicly confirmed this. Many believe that Taiwan Semiconductor Manufacturing Company, which makes chips for Nvidia, will not be able to supply the Santa Clara, California-based designer of GPUs enough 28nm products this calendar years.
Nvidia's Kepler family of products, which will likely get GeForce 600-series name in the consumer market segment, will not only power Nvidia's mid-term future products, but will also help Nvidia to boost sales of its desktop discrete graphics cards. In Q2 2011 shipments of discrete graphics boards for desktops were down 15%, according to some analysts.