IT News & Technology update

NVIDIA announces contest for CUDA developers

Written by IT News on 10:05 PM

NVIDIA is a waste of time to try more momentum for its line of high performance graphics processors and CUDA technology acquisition. The combination of the two solutions are designed to improve performance in applications that normally rely on CPU performance. For this reason announced graphics chip Maker recently that it will work with TopCoder, a competitive software development community, CUDA Superhero Challenge. This event is designed as a
series of competitions for programmers who will benefit from parallel processing with NVIDIA's CUDA architecture to solve some of the biggest challenges computer.

"PC architecture has evolved from only the central processing unit CPU co-processing on the CPU and GPU," said Sanford Russell, head of the group for NVIDIA CUDA. "By using the TopCoder community, we train over 200,000 programmers in parallel programming and offers advantages in blow GPU computing revolution. "

According to nVidia, will start the first match on 14 September and ends on 25 September. It is open to all eligible developers who participate in U.S. $ 5000 worth of prizes, which will be awarded to five winners. The company says that the contest winners will be announced on NVIDIA GPU Technology Conference, held in San Jose, California, from 30 September to 2 October 2009.

"We have the largest competitive software development community with more than 200 countries, and they are eager to show their talents in one of the fastest growing areas in the planning industry," said Rob Hughes, President and Chief Operating Officer at TopCoder. "We expect CUDA Superhero Challenge many creative approaches to solving computer will create challenges and will result in game changing innovations. "

No Comment

| Posted in » CUDA, CUDA developers, GPU, graphics, Nvidia

NVIDIA said to be working on enabling CUDA on other platforms

Written by IT News on 8:11 AM

NVIDIA said that work on the prestigious CUDA technology AMD's GPUs. This solution makes it possible for developers to optimize their products for use on GPUs from both NVIDIA and AMD with CUDA technology. Consequently, this will enable NVIDIA, currently the world's leading supplier of graphics processing units, to take advantage of GPU sales of its main competitor, AMD.

There are a number of general purpose GPU (GPGPU) computing standards that exist today, including Microsoft's DirectX and OpenCL Compute, both of which are currently supported by two leading graphics chip makers, NVIDIA and AMD. But the two competing companies which have their own solutions, NVIDIA placing a large bet on his CUDA technology and AMD is trying to lose space with its Stream solution. But according to a recent news article on the tech radar, where NVIDIA's Chief Scientist, Bill Dally, the company could be worked on further enlargement of the working environment for their CUDA solution.

"In the future, you can run CUDA C with an extension of a larger number of platforms, so I do not think it will be a fundamental limitation." Dally said. Although not directly confirming the support of AMD's platforms, he added, "I know a number of projects have been launched to CUDA on other platforms."

NVIDIA has an understanding for many applications, a statement which can be checked by taking a look at the hundreds of applications that are ported to CUDA, available at the specially designed CUDA zone site. Look at a comparison with AMD's solutions for NVIDIA's strong interest in the rapidly growing segment where AMD still lost its competitor.

"There are only a handful of things, and they are usually things that are already in CUDA. CUDA is much more get traction among the people programming parallel in many application. There is a simpler language to use," Dally stated.

No Comment

| Posted in » Amd, CUDA, GeForce, Nvidia, Stream

NVIDIA recongnizes Harvard as a CUDA Center of Excellence

Written by IT News on 1:18 AM

Santa Clara, California-based NVIDIA announced yesterday that Harvard University was a CUDA Center of Excellence. After the graphics chip manufacturers, recognition is justified as the University's commitment to teaching GPU computing, and integration of CUDA-capable GPUs for a variety of scientific and technological research projects. Before CUDA Center of Excellence Award, Harvard University a $2M grant from the National Science Foundation (NSF) for the development of GPU-based Computational Science.

With the interest in CUDA architecture quickly spread on campus at Harvard and scientific landscape living in Boston, it has never been a better time to announce this cooperation, "said Hans Peter Pfister, Gordon McKay Professor of Computer Science in practice is Harvard School of Engineering and Applied Sciences and Head of Visual Computing at the Harvard Initiative in Innovative Computing.”This generous gift from NVIDIA offers excellent learning opportunities for Harvard students, acceleration of our research and expands the use of GPUs for computer science and other demanding applications."

Harvard University already with the Graphics Processing Unit to provide solutions for a range of issues. For example, in a project called Connectome project, GPU is used to decode the complex structure of the human brain, a search that is a priority for the scientific community for a long time. While another project using the computing power of GPU is MWA telescope project, which is to discover the origins of the universe. Besides the above, the GPU is used for other power projects, as well as at Harvard University.

Back in mid 2008, the first NVIDIA CUDA Center of Excellence, an honor awarded at the University of Illinois at Urbana-Champaign, which includes a donation of $500,000 for the development of parallel computing and the continuation of research programs.

No Comment

| Posted in » CUDA, CUDA Center of Excellence, GPU computing, Harvard University, Nvidia

NVIDIA updated CUDA to 2.1 beta

Written by IT News on 3:14 AM

NVIDIA has been reported recently released a new version of its CUDA platform, a technology that is known to take advantage of GPGPU capacity graphic solutions company. Release does not come with new features, but the SDK has been added support for Windows Vista, Red Hat Linux, and Visual Studio 2008 in Windows XP.

The newly announced OpenCL standard set to make changes in GPGPU faster computers landscape, but the green it does not seem willing to let its CUDA platform to the side, even if it is based on a proprietary method to unlock the power of the graphics card maker's chips.

Version 2.1 beta of CUDA is offered as a free download, and it is that comes with Windows Vista support for Tesla products and 32-bit debugger support for CUDA on RedHat Enterprise Linux 5.x, as a separate download. This released also support Visual Studio 2008 on Windows XP and Vista, and Just-In-Time (JIT) compilation of applications, which dynamically generate CUDA kernels.

NVIDIA also said to have taken new interoperability API for Direct3D 9 and Direct3D 10 meant to speed up communication with DirectX applications. The OpenGL interoperability benefits of a series of improvements. Moreover, support for newer versions of Linux, including Fedora 9, OpenSuse 11 and Ubuntu 8.04, have been highlighted in the new version.

Santa Clara graphics card maker has recently announced its SDK can be used by developers to create platform-independent OpenCL-based GPGPU applications. For ordinary GPGPU software, proprietary developer’s solutions CUDA left some form 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, but there are some voices suggesting that the x 86 platforms cannot be started at all.

Currently, the only convincing solution development in the area seems to be OpenCL an application with the support not only on NVIDIA and AMD graphic solutions, but Intel's upcoming Larrabee graphics accelerator also.

The CUDA SDK 2.1 beta is available for download. The new CUDA toolkit is available on NVIDIA's website.

[source softpedia]