NVIDIA nTeresting News: 26 March 2010 |
Written by NVIDIA - Brian Burke | ||
Friday, 26 March 2010 | ||
NVIDIA nTeresting News: 26 March 2010In this Issue:
GeForce GTX 480 Comin'!The wait is over. Tonight's the big night at PAX East in Boston, MA! NVIDIA is at PAX in full force with:
If you are at the show, be sure and grab a seat at the Main Theater for the 6 to 7 p.m. stage show. You won't want to miss it. Not at PAX? Log into nvidia.com and check out the live blog from the event! Also, many websites will have reviews, so be on the look-out for those. CUDA Gets Re-uppedThis week we will release the CUDA Toolkit 3.0. Key features in this release include support for new GPUs based on Fermi architecture, C++ support and Improved developer tools for Linux. According to BSN* this will extend our GPU computing leadership" "Even though the world has yet to see an OpenCL application that makes any market impact, it looks like NVIDIA has raised the bar again. AMD's yet unannounced reorganization should help AMD to become more competitive in this field. From the other side, Intel currently isn't competing at all. All in all, CUDA Toolkit 3.0 looks like a very comprehensive suite for development of GPU-accelerated applications. We'll see how the competition will react to this resourceful toolkit." We support open standards, no question about it. But we also support standards that allow NVIDIA to innovate in a timely fashion, the way CUDA does. Developing new standards allows us to move features to GPUs as fast as possible. Our history with OpenCL illustrates our level of support is excellent:
• NVIDIA had the 1st OpenCL driver for the GPU.
• NVIDIA's Neal Trevett is the chairman of the Khronos Group that over sees OpenCL.
• We were first to submit a driver to the Khronos Group.
• We were first to give a driver to developers.
• We were first to demo OpenCL on a GPU.
• NVIDIA is the 1st and only company to provide a Visual profiler for development of OpenCL programs.
• NVIDIA is the 1st and only company to provide an OpenCL SDK for GPU's.
• NVIDIA is the 1st and only company to provide a best practices guide for OpenCL Programmers
Check out our nTersect blog post on the new CUDA toolkit for more details. AMD is Late to IE9 PartyLast week Microsoft introduced the BETA version of Internet Explorer 9. The big news was that Internet Explorer 9 will be GPU accelerated. NVIDIA blogged about it and we even had a quote in the Microsoft news release: "Internet Explorer 9 enabling GPU-accelerated HTML5 is a milestone for visual computing," said Drew Henry, general manager of GeForce and ION GPU business unit at NVIDIA Corp. "By harnessing the power of NVIDIA GPUs, Internet Explorer 9 removes the glass ceiling for Web developers, enabling them to build graphically rich, high-performing Web applications." This week AMD got interested and did a ‘me too' style blog post. "At last week's MIX10 event, several people talked about NVIDIA driving IE9 with GPU acceleration, but ATI was there as well. A new blog post from AMD reaffirms their own commitment to GPGPU, in particular the new IE9 features." Welcome to the party. Tegra Takes India by StormLast week we teamed up with partners EAFT, Compal and Hungama to showcase the latest devices and content based on our next-generation Tegra mobile technology coming onto the Indian market. Looks like the combination of top-quality multimedia, long battery life and Flash support has hit the spot. "At a press conference in Mumbai today, NVIDIA showcased two tablets running on the Tegra 2 chipset - Compal and Magic Tile. A horde of journalists descended upon the devices, photographing every conceivable angle and aspect of it."- Technoholik "The new Tegra boasts tremendous battery life which is possible with 8 different processors housed on the SoC that starts or stops according to the user's needs. Tegra 2 can give up to 16 hours of full HD video playback or up to 140 hours of music on a single charge." - TechTicker You can also check out ThinkDigit's video here. 3D Vision is the Best3D is hot. It was the hot topic at Tech Retreat this week. MaximumPC dig into this by publishing a Stereoscopic Imagine white paper in the latest issue. "After spending many hours comparing 3D-enabled games on a high-refresh-rate monitor and NVIDIA's LCD shutter glasses versus Zalman's Trimon monitor and circularly polarized passive spectacles, we developed a strong preference for NVIDIA's active solution." -MaximimPC, April 2010 page 60 NVIDIA 3D Vision is a combination of wireless active shutter glasses and advanced software, which automatically transforms hundreds of PC games into full stereoscopic 3D. In addition, you can view movies and digital photographs in eye-popping 3D. Adobe <3 GPUsAdobe has a great site that counts down the release of Creative Suite 5. As part of it they created a sweet (get it?) video that highlights how CUDA improves their Mercury playback engine. Essentially, Adobe has redesigned their entire video rendering and playback engine to harness the NVIDIA CUDA parallel processing architecture of Quadro graphics processing units (GPUs). The result is a fluid, real-time editing experience for adding additional effects, multiple layers, or ultra high resolution content. No more wasting time waiting for things like Encoding and Exporting progress bars to slowly fill the box. CUDA is Big in Japan's Weather BureauTiTech is using NVIDIA CUDA to achieve GPU acceleration for their next generation weather model developed by Japanese weather bureau. It scales 80 times faster than CPU. "Compared with CPUs, GPUs have very high processing performance and high computing performance per unit of power consumption. For this reason, 680 GPUs are used in the TSUBAME supercomputer at Tokyo Institute of Technology's Global Scientific Information and Computing Center. Because GPGPUs can utilize the video cards used in PCs, they also have the advantage of having very low cost." Professor Takayuki Aoki's research group at Tokyo Institute of Technology has achieved a performance of 44.3 GFlops (about 80 times of one CPU core) in GPU processing using the TSUBAME supercomputer, by migrating the next-generation weather model ASUCA entirely to GPU. Moreover, the group has achieved an execution performance of 3.22 TFlops by utilizing 120 GPUs connected via a high-speed network." CUDA is here today and is making a huge performance difference by giving developers a language and a great set of tools so their applications can take advantage of the massively parallel architecture of the GPU as a computational engine. |