CUDA version 1.1 is now released.
The toolkit and sdk have been updated since the beta release, so please install these newer versions.
There is a new CUDA Zone on the NVIDIA web site [url=“http://www.nvidia.com/object/cuda_home.html”]http://www.nvidia.com/object/cuda_home.html[/url]
The downloads are under “Get CUDA”
There is are also new ForceWare Release 169.21 Windows drivers available today that should be used with CUDA on Windows
Windows XP driver [url=“http://www.nvidia.com/object/winxp_169.21.html”]http://www.nvidia.com/object/winxp_169.21.html[/url]
Windows XP x64 driver [url=“http://www.nvidia.com/object/winxp64_169.21.html”]http://www.nvidia.com/object/winxp64_169.21.html[/url]
The 169.21 driver adds support for the new G92-based GeForce 8800 GTS 512. CUDA 1.1 works fine with this card (~300 GFLOPS nbody, ~350 GFLOPS VMD)
The MATLAB examples have been updated for 1.1. There is a new FORTRAN example using pinned memory with CUBLAS. See the links under Developing with CUDA on the CUDA Zone.
There is also an initial list of CUDA applications & papers being populated under the CUDA Showcase section. If you would like to get added or need anything changed please send details to cuda@nvidia.com (sorry if we missed you… somehow Manifold System, PyStream, SnapCT and HOOMD didn’t even make this first round)
The CUDA Rocks Roll for clusters has not yet been updated for CUDA v1.1