by Anton Shilov
12/13/2012 | 01:26 PM
Nvidia Corp. continues to reveal additional details regarding its highly-anticipated project code-named Denver, a general-purpose fully-custom ARMv8 architecture compliant microprocessor core with a “secret sauce”. As it appears, despite all the hype around micro-servers these days, Nvidia will not make a special chip for that part of the market and will position Denver for high-performance computing instead.
“Right now we are very focused on high-performance computing. Strategy for Denver, which is a fully-custom ARM core […], is to find a way into Tesla applications. It makes a lot of sense on the server side to put it into Tesla so that to have a differentiation on the HPC market. Today we are not looking to attack the general-purpose server market, that market is looking quite crowded […] it will be a tough place to be. So, we will focus on the market we already know […] and which is a multi-billion dollar opportunity,” said Chris Evenden, director of investor relations Nvidia, at Raymond James IT supply chain conference.
As is known, Denver will find home inside next-generation Maxwell graphics processor, therefore, will not quite suite for ultra low-power or general-purpose servers, unless they will require extreme performance in data crunching applications.
It is rumoured that Nvidia is also developing high-end fully-custom ARM core code-named Boulder. That chip is supposed to compete for general-purpose servers, or for whatever thing they develop into within the next several years. While Mr. Evenden did not confirm the existence of the project Boulder, he said “never say never”, when asked about possible rivals for AMD Opteron and Intel Xeon chips.
Known under the internal code-name "project Denver”, the Nvidia initiative includes an Nvidia CPU running the ARM instruction set, which will be fully integrated on the same chip as the Nvidia GPU. The hybrid processors which will have both custom 64-bit ARMv8-compatible general-purpose cores as well as Nvidia's custom compute cores known as stream processors will be aimed at various market segments, most importantly at high-performance computing. Since technologies developed within project Denver are universal, they will eventually span across the whole lineup of Nvidia products, from Tegra to GeForce to Quadro to Tesla. Obviously, Denver-derivatives may power next-generation game consoles as well.
Recently Nvidia's chief executive did not exclude the Denver-based chips from entering very energy-efficient high throughput servers.