Nvidia Makes Fermi Mobile

Nvidia has introduced new graphics cards for mobile workstations with Fermi architecture. And promptly doubled the RAM and CUDA cores.

Archive notice: This article was originally published on February 24, 2011. Links and embedded videos are preserved as part of the historical record.

For cutters, editors and engineers, Quadro GPUs with Fermi architecture are now also available in mobile workstations. In contrast to the previous series, CUDAs and RAM have been doubled. To conserve the battery (although hardly anyone works on the road using only the battery), “Optimus” technology makes energy consumption more efficient: cores are switched on only as needed, preventing the power store from being drained outside rendering times.

The flagship Quadro 5010M brings a hefty 4GB of GDDR5 memory, which calculates almost a billion triangles per second. As can be read in the new issue of DP, the CUDA processor architecture brings performance increases in compute-intensive applications such as ray tracing, video editing and scientific applications. The GPU Tessellation Engine is now also mobile; it supports OpenGL 4.1 and Shader Model 5.0 and enables extremely detailed geometry for models and environments when creating film scenes. Microsoft DirectX 11 and the DirectCompute and OpenCL standards are also included.

The new models are:

Quadro 5010M (flagship): 384 CUDA cores; four GB GDDR5 memory; designed for new mobile workstations with a 17.3-inch display

Quadro 4000M (production class): 336 CUDA cores; two GB GDDR5 memory; designed for new mobile workstations with a 17.3-inch display

Quadro 3000M (production class): 240 CUDA cores; two GB GDDR5 memory; designed for new mobile workstations with a 17.3-inch display

Quadro 2000M (mid-range): 192 CUDA cores; two GB DDR3 memory; designed for mobile workstations with a 15.6-inch display

Quadro 1000M (mid-range): 96 CUDA cores; two GB DDR3 memory; designed for mobile workstations with a 15.6-inch display.