In the ongoing battle between OpenCL & CUDA, AMD has launched the next volley with their latest AMD Accelerated Parallel PRocessing SDK v2.5.
In related news, AMD earlier this month announced [press release] the availability of its AMD Accelerated Parallel Processing (APP) Software Development Kit (SDK) v2.5. This OpenCL-driven SDK offers programmers tools to add general-purpose GPU (GPGPU) computing support to their applications. AMD seems to be taking a wise tact here, as its CPU performance trails that of rival Intel Corp. (INTC), but its GPU performance is well ahead of its rival. With GPU-enabled apps, Fusion APUs may finally start performing their Intel comparables in everyday applications like Microsoft Corp.’s (MSFT) Office suite or Adobe Systems, Inc.’s (ADBE), assuming the app-makers add support.
This comes along with the announcement of their new C-series and E-series processors, boosting the speeds of their existing Fusion APU’s to new heights. The new E-series is, however, a bit slower and seems to be targeted more at the ultra low-power mobile market. Both series however now support DisplayPort++ and HDMI 1.4a.
VisualisingData has a great list of visualization tools, ranging from the big dogs (R, Processing) down to the lesser-known guys (Mondrian, D3) that winds up being a great starting list for anyone looking to get into the data visualization space.
This second part presents the prominent and powerful data visualisation programming languages and environments that dominate the creative engineering that sits behind today’s design output
Realizing the success of Nvidia’s CUDA university initiatives, AMD recently announced a new OpenCL University Kit, a collection of materials that can be used in any university environment to teach OpenCL programming.
“Teaching students to effectively leverage the OpenCL standard involves all the intricacies of parallel programming plus support for a new class of heterogeneous computing devices built on a variety of hardware technologies,” said David Kaeli, professor and associate dean of undergraduate programs, Northeastern University College of Engineering. “The OpenCL University Kit introduced by AMD is an easy tool to enable educators to quickly introduce OpenCL learning into their curriculum, helping them strike a balance between teaching syntax and higher level architectural issues.”
The kit includes 13 lectures, with instructor and speaker notes, as well as code examples. Combined with the recently announced ‘Accelerated Parallel Processing SDK‘, (the new name for the old Stream SDK) it’s a great way to get into OpenCL development.
Qualcomm has announced the winners of their 2010 Augmented Reality Developer Challenge. A clever way to push usage of their new AR SDK, I see some pretty clever uses of the technology for some fun little games.
Winning Applications
1st Place – $125,000 – Paparazzi by Paulius Liekis and Arminas Didžiokas (Lithuania)
An interactive game where the player becomes a virtual paparazzo and sneaks pictures of a vain celebrity before he gets agitated and attacks the photographer
2nd Place – $50,000 – Inch High Stunt Guy by Defiant Development Pty Ltd. (Australia)
A game where the player arranges various obstacles to enable a stuntman to successfully jump his motorcycle through a hoop
3rd Place – $25,000 – Danger Copter by Alex Beachum, Jonghwa Kim, Jason Mathias, Kedar Reddy and Evan Sforza (USA)
A gaming adventure where the player becomes a helicopter pilot who maneuvers a water-spouting chopper to extinguish fires and rescue people from danger
Congratulations to all the winners! Check out the “sizzle reel” below which shows some of the better entries.
A new competitor to SpeedTree is on the horizon in a tool called “Woody3D”. Fully animated tree generation and rendering with just a click of a button, and it now comes complete with C++ and SDK code to allow you to link it into any application of your choosing. The new version even comes with a free license allowing you to “try before you buy”, really giving you no reason not to at least give it a shot.
If you decide to buy, it’s available for $99 with a license that allows 2 PC’s per seat, and an unlimited number of applications (Although a commercial license is required for apps grossing over $250k)
Hit their website for all the details and downloads.
Big news for Visual Studio developers using CUDA, NVidia has just announced that the newest version of PArallel NSight 1.51 Professional Edition is now available for free for all! Coming with all of the great professional tools, this is huge news and a mandatory download for anyone doing NVidia GPU development on Windows.
Parallel Nsight Professional Edition is now available for all Visual Studio 2008 and 2010 developers, free of charge. NVIDIA is now offering the full Parallel Nsight Pro feature set at no cost, as we historically have done with CUDA and our other development tools, so that a broader range of developers can take advantage of the full benefits of this popular parallel computing development tool. Parallel Nsight support will continue via the Parallel Nsight forums, and Professional developers are encouraged to sign up for NVIDIA’s Registered Developer Program, which provides priority access to new software, bug management tools, invitations to members-only developer webinars and other development resources.
In addition to making it a free download, they’ve removed all of the license key and activation stuff, so it really is a ‘no strings attached’ release.
Equalizer, the somewhat next-generation of the massive tiled display system Chromium, has just hit version 1.0 alpha and added a nice API and SDK system called ‘Collage’.
The most notable new features in this release are:
Full feature set and API of Equalizer 1.0
GPU compression plugins
Failure tolerance during initialization
Administrative API for runtime configuration changes (preview)
Intended primarily for application developers, it’s a great API for creating massively parallel OpenGL applications.
In an attempt to stem the many photoshopped images that wind up on the front pages of major newspapers, Canon has integrated a special crytographic security measure that allows someone to determine if the image is an original or has been altered.
In brief, modern DSLR (Digital Single-Lens Reflex) cameras produced by Canon feature Original Data Security system which is meant to securely validate the authenticity of image data and prove image genuineness. Accordingly, one can use OSK-E3 (Canon Original Data Security Kit) which comprises smart card and special software to verify a digitally signed image.
Unfortunately, ElcomSoft today revealed a vulnerability in their algorithm that allows anyone to cryptographically sign any image so that it appears authentic.
ElcomSoft discovered the vulnerability which allows producing images that will be positively validated by Canon’s own Original Data Security Kit (OSK-E3) regardless of whether or not the images are, in fact, genuine.
See some humorous images on their site, as well as the PDF detailing the vulnerability.
Yesterday NVidia announced the newest release of the CUDA Toolkit, version 3.2, that not only offers some new CUDA libraries for things like sparse matrix work and random number generation (important for crytography), but a nice 300% performance boost in the older CUDA BLAS libraries. They’ve also added an h.264 encode and decode algorithm to the mix, and some cluster management tools for folks running bigger GPGPU setups.
If you want to know more, they are also hosting a free webinar Tuesday November 23rd at 10am PT to discuss it. Get all the details in the press release after the break.
Possibly to compete against NVidia & PGI’s ‘CUDA x86‘ offering, Intel has announced at SC10 that they have a new OpenCL SDK enabling executing of OpenCL code on Intel x86 Processors, currently only on Windows Vista & Windows 7.
OpenCL* is an emerging standard from the Khronos Group industry consortium. As a Khronos founder and promoter, Intel has made significant contributions to OpenCL* feature set. With the Alpha release of Intel® OpenCL SDK, Intel continues to demonstrate its commitment to parallel computing tools and standards support.
OpenCL still hasn’t achieved the penetration of CUDA, primarily because it’s a bit more difficult to work with and it simply hasn’t had the time that CUDA has had, but this is a huge step toward creating a single truly universal language that runs on AMD GPU’s, NVidia GPU’s, and now Intel CPU’s.
In the ongoing battle between OpenCL & CUDA, AMD has launched the next volley with their latest AMD Accelerated Parallel PRocessing SDK v2.5.
VisualisingData has a great list of visualization tools, ranging from the big dogs (R, Processing) down to the lesser-known guys (Mondrian, D3) that winds up being a great starting list for anyone looking to get into the data visualization space.
Realizing the success of Nvidia’s CUDA university initiatives, AMD recently announced a new OpenCL University Kit, a collection of materials that can be used in any university environment to teach OpenCL programming.
A new competitor to SpeedTree is on the horizon in a tool called “Woody3D”. Fully animated tree generation and rendering with just a click of a button, and it now comes complete with C++ and SDK code to allow you to link it into any application of your choosing. The new version even comes with a free license allowing you to “try before you buy”, really giving you no reason not to at least give it a shot.
Big news for Visual Studio developers using CUDA, NVidia has just announced that the newest version of PArallel NSight 1.51 Professional Edition is now available for free for all! Coming with all of the great professional tools, this is huge news and a mandatory download for anyone doing NVidia GPU development on Windows.
Equalizer, the somewhat next-generation of the massive tiled display system 
In an attempt to stem the many photoshopped images that wind up on the front pages of major newspapers, Canon has integrated a special crytographic security measure that allows someone to determine if the image is an original or has been altered.
Yesterday NVidia announced the newest release of the CUDA Toolkit, version 3.2, that not only offers some new CUDA libraries for things like sparse matrix work and random number generation (important for crytography), but a nice 300% performance boost in the older CUDA BLAS libraries. They’ve also added an h.264 encode and decode algorithm to the mix, and some cluster management tools for folks running bigger GPGPU setups.
Possibly to compete against NVidia & PGI’s ‘
VizWorld.com 
Comments