Reporting on overheard information from a number of unnamed sources, Theo Valich claimed today that NVIDIA has embarked on finding design targets for their upcoming Project Denver release.
Although the writer reminds readers to take the news with a grain of salt given the unverified sources, the Project Denver CPU core is “looking to be very much aligned with T40 i.e. ‘Tegra 4’ i.e. Wayne and, according to schedule, Wayne silicon is going to be taped out in the next couple of weeks.” The rumor is that developers will be able to get their first view of the 4-ARM core prototype for Christmas.
If Valich’s sources are correct, December will be the month NVIDIA tapes out the first silicon that is based on the Denver design which blends up to 8-core custom NVIDIA-ARM 64 CPUs with NVIDIA’s GeForce 600-class GPU. He notes that the company has hit some roadblocks during the CPU side of development and that they are taking a more cautious approach by sticking with their Fermi roots in the design.
He writes that another source that is close to the development process claimed that the filling for the GPU will consist of at least 256 cores. Valich goes on to write that this:
“…would put the product on pair with AMD’s Trinity APU which will pair a Bulldozer-Enhanced CPU core with “Northern Islands” VLIW4 architecture and will be the key APU for AMD in 2012. Compute power-wise, NVIDIA doesn’t want to clock it to heavens’ high, but rather to squeeze each IPC (Instruction Per Clock) as possible. Still, it is realistic to expect 2.0-2.5GHz for CPU and similar clock for the GPU part, with memory controller and the rest of the silicon working at a lower rate to keep everything well fed.”
He claims that December could prime NVIDIA for the blade market with its ability to pull a robust GPU into the market, exterminating the requests for Intel or AMD x86 cores for their Tesla market.