Nvidia has unveiled its CUDA GPU architecture codenamed ‘Fermi’, which it claims to accelerate performance on a wider array of computational applications.
According to Nvidia, as the foundation for its family of GPUs namely GeForce, Quadro and Tesla ‘Fermi’ features a host of new technologies including, C++, which complemets existing support for C, Fortran, Java, Python, OpenCL and DirectCompute; ECC, which is required in datacenters and supercomputing centers deploying GPUs on a large scale; 512 CUDA Cores that feature IEEE 754-2008 floating-point standard; and 8x the peak double precision arithmetic performance, for high-performance computing (HPC) applications.
Other features include Parallel DataCache that speeds up algorithms such as physics solvers, raytracing, and sparse matrix multiplication; GigaThread Engine with concurrent kernel execution, where different kernels of the same application context execute on the GPU at the same time; and an integrated computing application environment Nexus.
Jen-Hsun Huang, co-founder and CEO of NVIDIA, said: It is completely clear that GPUs are now general purpose parallel computing processors with amazing graphics, and not just graphics chips anymore. The Fermi architecture, the integrated tools, libraries and engines are the direct results of the insights we have gained from working with thousands of CUDA developers around the world. We will look back in the coming years and see that Fermi started the new GPU industry.
