NVidia has released a plug-in example that allows Matlab programs to use standard graphics processing unit (GPU) libraries to speed up their applications. The plug-in example will also show users how to write their own versions, enabling them to take the performance critical piece of their code and harness the capabilities of the GPU through the NVidia Cuda software environment.
A typical Matlab simulation of 2D isotropic turbulence at a resolution suitable for scientific publications (1024x1024) would typically take a couple of days, but with the plug-in the simulation takes as little as four hours.
The announcement comes soon after the company’s introduction of the NVidia Tesla family of GPU computing solutions last week, and the release of NVidia Cuda 1.0, the production release of the C-compiler and Software Development Kit (SDK) for developing computing applications on NVidia GPUs.
The combination of GPU computing technology and the Cuda software environment delivers a flexible, massively parallel computing platform for today’s most demanding data-intensive applications.
Cuda 1.0 includes new C compiler optimisations and performance enhancements along with additional functionality and C code examples. Cuda Blas and FFT libraries have been further optimised and include additional functionality and new C code examples relevant to areas such as computational finance and medical imaging are installed with the SDK.