NVidia Announces New Kepler-Based GPUs

NVidia's new Tesla GPUs surpass its predecessor, NVidia Fermi.

NVidia has announced a new family of Tesla GPUs based on the Kepler GPU computing architecture. According to the company, the new NVidia Tesla K10 and K20 GPUs are computing accelerators built to handle the most complex high-performance computing problems in the world. According to the company, the Kepler architecture was designed with an intense focus on high performance and extreme power efficiency and is three times as efficient as its predecessor, the NVidia Fermi architecture. Click here for more information. http://www.nvidia.com/object/tesla-supercomputing-solutions.html

According to the company, the NVidia Tesla K10 GPU delivers the world's highest throughput for signal, image, and seismic processing applications. Optimized for customers in oil and gas exploration and the defense industry, a single Tesla K10 accelerator board features two GK104 Kepler GPUs that deliver an aggregate performance of 4.58 teraflops of peak single-precision floating point and 320GB per second memory bandwidth.

In addition to the Kepler architecture, NVidia released a preview of the CUDA 5 parallel programming platform. The platform will enable developers to begin exploring ways to take advantage of the new Kepler GPUs, including dynamic parallelism. NVidia also worked with LLVM developers to provide its CUDA compiler source code changes to the LLVM core and parallel thread execution back end. LLVM is a widely used open source compiler infrastructure with a modular design, and the CUDA compiler provides C, C++, and Fortran support for accelerating application using NVidia's GPUs.  

The latest version of the LLVM compiler with NVidia GPU support can be downloaded through the LLVM site.