This Article 
 Bibliographic References 
 Add to: 
Fermi GF100 GPU Architecture
March/April 2011 (vol. 31 no. 2)
pp. 50-59

The Fermi GF100 is a GPU architecture that provides several new capabilities beyond the Nvidia GT200 or Tesla architecture. The Fermi architecture offers up to 512 CUDA cores and special features for gaming and high-performance computing. This article describes the GPU's new capabilities for tessellation, physics processing, and computational graphics.

1. NVIDIA, "Fermi: NVIDIA's Next Generation CUDA Compute Architecture," 2009; NVIDIA_Fermi_Compute_Architecture_Whitepaper.pdf .
2. C.M. Wittenbrink, E. Kilgariff, and A. Prabhu, "Fermi GF100: A Graphics Processing Unit (GPU) Architecture for Compute Tessellation, Physics, and Computational Graphics," IEEE Hot Chips, presentation, 2010; HC22.23.110-1-Wittenbrink-Fermi-GF100.pdf .
3. T. Purcell, "Fast Tessellated Rendering on the Fermi GF100," High Performance Graphics Conf., Hot 3D presentation, 2010; www_2010/media/Hot3DHPG2010_Hot3D_NVIDIA.pdf .
4. J. Nickolls and W.J. Dally, "The GPU Computing ERA," IEEE Micro, vol. 30, no. 2, 2010, pp. 56-69.
5. NVIDIA, "NVIDIA GF100: World's Fastest GPU Delivering Great Gaming Performance with True Geometric Realism," 2010; .
6. Microsoft, "MSDN Tessellation Overview" 2010; .

Index Terms:
microarchitecture, graphics processing unit, Nvidia, Fermi GF100, GT200, Tesla, tessellation, physics, computational graphics
Craig M. Wittenbrink, Emmett Kilgariff, Arjun Prabhu, "Fermi GF100 GPU Architecture," IEEE Micro, vol. 31, no. 2, pp. 50-59, March-April 2011, doi:10.1109/MM.2011.24
Usage of this product signifies your acceptance of the Terms of Use.