| NVIDIA nTeresting 23 October 2009 |
| Written by NVIDIA - Brian Burke | |||
| Friday, 23 October 2009 | |||
NVIDIA nTeresting 23 October 2009In this issue:
3D Power From the CloudsNVIDIA and mental images introduced the NVIDIA RealityServer at the Web 2.0 Summit in San Francisco. "Computer graphics usually comes with a tradeoff: Users get to see extremely realistic images, or pictures that can be viewed interactively, but not both. Nvidia believes those days are ending. The Silicon Valley chip company on Tuesday announced plans to offer a combination of hardware and software that can generate three-dimensional images that are almost indistinguishable from photographs-and do so in a matter of seconds, not the hours that such chores typically require." NVIDIA RealityServer is a cloud computing solution, and the first ever 3D Web services platform powered by GPUs. It is powerful combination of NVIDIA GPUs and mental images software that streams interactive, photorealistic 3D applications to any web connected PC, laptop, netbook and smart phone. It is comprised of NVIDIA Tesla RS servers, RealityServer 3.0 software from mental images, and features iray interactive ray tracing technology. DirectCompute Means GPU Computing is Here to StayWe've already seen some applications come to market that take advantage of the GPU and they have shown dramatic improvements in performance. DirectCompute will make it even easier for developers to write applications and deliver these benefits to millions of additional customers and turbo-charge digital media applications. We believe DirectCompute will take GPU computing from a niche to the mainstream by making this technology available to the millions of users of the Windows 7 operating system. So does Extremetech. "It certainly means that in the software world, writing certain kinds of applications to use the GPU (as well as to better use multiple CPU cores) is becoming much more commonplace. Whether OpenCL, DirectCompute, or a GPU-specific development kit, GPU computing will be part of computing going forward." They also know who to credit for the GPU computing revolution. "NVIDIA, as the largest company that focuses on graphics processing, has been the most visible company promoting the concept of using the GPU for computing. It has created CUDA architecture, essentially a framework for developers to use to enable their programs for parallel processing on NVIDIA's current and upcoming graphics cards, as well as on Tesla, special-purpose cards with high-end GPUs designed for using CUDA code in high-performance computing applications." We have demonstrated our support for DirectCompute by being the first to release support for DirectCompute in our GPUs. We support open standards, plus standards that allow NVIDIA to offer new innovations to customers well in advance of industry standards, such as CUDA C and PhysX. Our goal is to lead the industry in new amazing directions and create value for our customers. GPUs Remake the Film industryAdding special effects in to films used to be a painstaking process that took weeks or months. Thanks to the GPU, that is changing. Leading industry players like Richard Kerris, chief technology officer of Lucasfilm, part of Industrial Light and Magic (ILM) are on board. "The advent of the GPU is really the next big frontier for us. We have seen hundreds of times improvements over the last few months. This is taking Moore's Law out the window," Mr Kerris told BBC News. Back in the day, the simplest of special effects rendering took a lot of computing power and a 500-square-foot room back then that really wouldn't operate our phone systems today. But, the talent and the understanding of what could be done out of that was able to produce movies like Terminator. It was cutting-edge stuff and getting a computer that was the size of a small automobile to render out simulations of a twister was pretty groundbreaking then," said Mr Kerris. NVIDIA's vision of the GPU is bigger than 15 or 20 more frames per second in a shoot ‘em up game. We want that, too, but we do not think we have to stop there. Fermi is the world's first computational GPU, that is, the first GPU designed from the ground up for compute. And it will kick ass in games, too. NVIDIA Recognized for HPCA panel of leaders in HPC and supercomputing have nominated NVIDIA for an award in the "Organization" category of insideHPC's 2009 HPC Community Leadership Awards. What is more impressive are our running mates - The DOE SciDAC Program, Oak Ridge National Lab and the Texas Advanced Computing Center. Needless to say, to be nominated amongst these pioneering organizations is a huge honor! Our Chipsets are Better than Intel'sNVIDIA makes great chipsets. Windows Middle East Magazine gave NVIDIA ION a Product of the Year Award. ION also impressed MobileTechReview: "If we gave Editor's Choice awards to technologies rather than products, we'd give one to NVIDIA for their ION platform. It will shift the netbook from web surfing and email machine to a movie and gaming portable." Our GeForce 9400M chip is also really good: "One of the highlights of the 14z is undoubtedly the NVIDIA GeForce 9400M graphics. Not only can they provide accelerated video decoding, but they end up being the best current IGP solution in terms of graphics performance -- roughly 50% faster than the ATI HD 3200 (although that's not an exact figure considering the different CPUs). That puts Intel in an interesting position for laptops, as the best combination right now appears to be an Intel CPU with an NVIDIA chipset..." Everyone loves our chipsets, well, everyone except Intel. "While Atom netbooks with Intel integrated graphics (that's every netbook shipped with the exception of the HP Mini 311 as of this writing) do a poor job of video playback, an ION-equipped netbook is something of an idiot savant-- it doesn't have the CPU to do heavy computational tasks but it can play video like a champ. It does a better job than full-sized notebooks with Intel integrated graphics, in fact." That should explain at lot. Just sayin'.... Notebook Hits Keep on Coming!Notebooks featuring NVIDIA GPUs continue to win awards. This time it is notebooks.com and the Alienware M17x, a multiple repeat offender when it comes to Editor's Choice Awards. "The M17x NVIDIA sent over for review is priced at $2,624, making it much more approachable for gaming enthusiasts. At this price, the M17x comes with dual NVIDIA GeForce GTX 280M GPUs with 2GB of memory, 4GB of RAM, 500GB 7200RPM hard drive and a DVD burner." Was it fast? "The system delivered awesome graphics performance even though it was paired with the most affordable processor offered with the M17x." That is an optimized PC. "Game play was very smooth and realistic, in part thanks to PhysX support. Here's a quick demo video that shows how the game looks on the M17x running at 1920×1200 with all of the effects turned on, detail level set to very high and Physx set to high." That is ‘Graphics Plus'. GeForce GPUs offer graphics performance that is second to none, and "Graphics Plus" features including:
Graphics on a BudgetThe GeForce 2200 GT continues to be called out for the value it provides in entry-level graphics. It is a nice bump up from integrated graphics for a low price. It offers Premium Windows 7 experience and great video options. "The card is very small, uses low power and will work great in any HTPC as it plays DVD and Blu-Ray without a hiccup." Diesel-Powered Supercomputing Needs GeForceBright Side of News has a story that results from a GTC keynote. When you can't plug in to an electrical socket, whatever processor you use for computing better be efficient. In Australia, they go off the grid and turn to a diesel-powered supercomputer. "In October 2006 NVIDIA launched CUDA, following up with GeForce 8 hardware in November 2006. According to scientists from this project, this was the beginning of a breakthrough, with GeForce GTX 280 and Tesla C1060 cards winning the computing challenge. In around 1kW of power, scientists managed to squeeze 4.5TFLOPS [dual GTX 295 card], meaning "only" 5.5kW is needed for 20TFLOPS." It is all about using the right tool for the job. If the job is massively parallel, that tool is a GPU. |
|||
