With the “Fusion” era beginning early 2011, accompanied by the release of Intel’s Sandy Bridge and AMD’s Fusion C-series and E-series, NVIDIA’s response has been anticipated. Both Intel and AMD are moving towards heterogeneous chips, combining CPU and GPU on the same die. Judging by NVIDIA’s partnership with ARM for Tegra, and CEO Jen-Hsun Huang’s quote “ARM is our CPU strategy”, an ARM based CPU with an NVIDIA GPU has been rumoured for a while. The same has now been confirmed, codenamed “Project Denver”.

With the “Fusion” era beginning early 2011, accompanied by the release
of Intel’s Sandy Bridge and AMD’s Fusion C-series and E-series, NVIDIA’s
response has been anticipated. Both Intel and AMD are moving towards
heterogeneous chips, combining CPU and GPU on the same die. Judging by
NVIDIA’s partnership with ARM for Tegra, and CEO Jen-Hsun Huang’s quote
“ARM is our CPU strategy”, an ARM based CPU with an NVIDIA GPU has been
rumoured for a while. The same has now been confirmed, codenamed
“Project Denver”.

In essence, Project Denver is Tegra super-sized. While Tegra was designed for smartphones with a tiny power-draw, Project Denver is designed for desktops, workstations, servers and even supercomputers. Project Denver includes a custom-designed ARM CPU, which is likely to be much higher performance with higher power draw than ARM’s familiar CPUs, fully integrated with an NVIDIA GPU on the same die. Essentially, NVIDIA’s Fusion.

However, it is not a direct competitor to Intel and AMD, using the ARM ISA over x86, which is the current standard in the PC world. NVIDIA believes ARM is the architecture of the future. On release, OS choices will range from Google’s Android, Symbian, iOS to the upcoming Windows ARM OS for low-power devices. How the PC landscape and software will support the ARM ISA for high performance and gaming applications remains to be seen.

NVIDIA has chosen the right time to announce Project Denver – right in the mix of Intel and AMD’s Fusion releases. No details or specifications are mentioned, however.

Source: NVIDIA Press Release