The new release features extended support on different OSes and APIs

Dec 22, 2008 08:33 GMT  ·  By

NVIDIA is reported to have recently released a fresh version of its CUDA platform, a technology known to benefit from the GPGPU capabilities of the company's graphics solutions. The release does not come with new features, but the SDK has been added support for Windows Vista, Red Hat Linux, as well as VisualStudio 2008 on Windows XP.

The newly announced OpenCL standard is set to make changes in the GPGPU accelerated computing landscape, yet the green company does not seem willing to let its CUDA platform aside, although it is based on a proprietary approach on unleashing the power of the graphics card maker's chips.

Version 2.1 beta of CUDA is offered as a free download, and it’s coming with Windows Vista support for Tesla products and 32-bit debugger support for CUDA on RedHat Enterprise Linux 5.x, as a separate download. The release also brings VisualStudio 2008 support on Windows XP and Vista and Just-In-Time (JIT) compilation for applications, which dynamically generate CUDA kernels.

NVIDIA also said having included new interoperability APIs for Direct3D 9 and Direct3D 10 meant to accelerate the communication with DirectX applications. The OpenGL interoperability also benefits from a series of improvements. In addition, support for recent releases of Linux, including Fedora 9, OpenSuSE 11 and Ubuntu 8.04, has been featured in the new version.

The Santa Clara graphics cards maker has recently announced that its SDK could be used by developers for the creation of cross-platform OpenCL-based GPGPU applications. For mainstream GPGPU software development, proprietary developer solutions such as CUDA remain some sort of a roadblock. As many of you already know, NVIDIA announced earlier this year its plans to release an x86 multi-core processor version of CUDA. The release has been delayed for the fourth quarter of this year, yet there are some voices suggesting that the x86 platform might not be launched at all.

Currently, the only convincing development solution in the area seems to be OpenCL, an application featuring support not only on NVIDIA or AMD graphics solutions, but on Intel's future Larrabee graphics accelerator as well.

The CUDA SDK 2.1 beta is available for download on Softpedia via this link. The new CUDA toolkit can be found on NVIDIA's site.