|
The ATI Stream Team Quarterly is a regular newsletter to keep you up-to-date about ATI Stream.
ATI Stream technology is a set of advanced hardware and software technologies that enable AMD graphics processors (GPUs), working in concert with the system's central processors (CPUs), to accelerate applications beyond traditional graphics and video processing. This helps PC's and servers to run computationally-intensive tasks more efficiently, providing a superior application experience to the end user. |
|
|
IN THIS ISSUE:
|
|
|
With the recent release of the OpenCLT GPU Beta as part of the ATI Stream SDK v2.0 Beta Program, this quarterly is packed full of OpenCL information and resources. If you have any suggestions about what you would like to see in the next newsletter, please do not hesitate to email us at: streamcomputing@amd.com.
ATI Stream technology is enjoying significant momentum. We look forward to bringing you even more success stories and product news on a regular basis.
Sincerely,
The ATI Stream Team |
|
| |
|
|
| |
 |
|
OpenCLT CPU+GPU Beta Release
AMD is proud to offer the fourth beta release of ATI Stream SDK v2.0, providing the first complete OpenCL development environment for GPUs and CPUs. This release is certified fully compliant with OpenCL 1.0 by the Khronos Group and is supported on a wide range of AMD GPUs, as well as any x86 multi-core CPU supporting SSE3 or greater. AMD delivers both high-performance CPU and GPU technology, giving us the unique ability to provide |
|
|
developers with an OpenCL platform that enables developers to leverage all the available processors in the system! The beta is available for immediate download as part of our ATI Stream SDK Beta Program and we encourage you to take a look.
Introductory Tutorial to OpenCLT with Benedict Gaster
Benedict Gaster, our OpenCL compiler architect here at AMD, has written an introductory tutorial for OpenCL to help developers get started learning and getting comfortable programming in OpenCL.
You can review his full OpenCL tutorial article.
AMD Developer Inside Track: Introduction to OpenCLT with Michael Houston
Michael Houston, a GPG system architect, gives developers an introduction to OpenCL. Learn about OpenCL, what the transition to OpenCL will be like, what applications are ideal for OpenCL and what impact OpenCL is expected to have on future software. View the video. |
|
| |
 |
|
Image Convolution Using OpenCLT - A Step-by-Step Tutorial
Udeepta Bordoloi, one of our ATI Stream application engineers, takes developers through a step-by-step tutorial illustrating techniques for implementing image convolution using OpenCL. This is a step-by-step tutorial series targeted at the beginner/intermediate level OpenCL user. He takes us through the initial OpenCL program, OpenCL with C++ bindings and a few optimizations. View tutorial. |
|
|
 |
|
OpenCLT Tutorial: N-Body Simulation
David Richie, from Brown Deer Technology, shows developers how he implemented a simple N-Body simulation in OpenCL. This algorithm is used frequently in demonstrations of computational performance and is an interesting algorithm because it represents a general class of algorithms with applications ranging from chemistry to astrophysics. View the tutorial. |
|
|
 |
|
Spotlight Application: Distributed RC5 Encryption with ATI Stream
Distributed.net has used AMD GPUs to accelerate the discovery of solutions for the RC5-72 secret key challenge. Using the ATI Stream SDK, the team has achieved a key discovery rate of over 1200 MKeys/sec using the latest ATI RadeonT HD 5870 GPU*, easily several times faster than the currently available cards from our competition. For more information, visit the ATI Stream Developer Showcase section on security applications and look for "RSA Labs RC5-72 Secret-Key Challenge Client". |
|
|
*Based on AMD internal testing using RC5-72 clients as of 9/04/09. Results shown in MKeys evaluated per second. Configuration: AMD PhenomT X4 9950 Black Edition processor, 8GB DDR2 RAM, Windows Vista® 32-bit. AMD drivers: ATI CatalystT 9.8 (ATI RadeonT HD 48xx), prerelease driver (ATI Radeon HD 5870). Nvidia driver: GeForce 190.62. AMD client: [x86/Stream], v2.9106.513 (beta8). Nvidia client: [x86/CUDA-2.2], v2.9105.512 (beta8).
AMD and SiSoftware Collaborate on OpenCLT Industry Benchmark Suite
AMD has worked with SiSoftware to develop one of the first industry benchmarks testing suites for OpenCL, which was released in the middle of November. The suite of OpenCL GPGPU benchmarks is part of SiSoftware Sandra 2010, the latest version of the award-winning utility first introduced in 1997, and includes remote analysis, benchmarking and diagnostic features for PCs, servers, mobile devices and networks. The OpenCL benchmark suite from SiSoftware is a first step in giving the industry the tools it needs to accurately measure and assess system performance in order to make decisions with confidence. For more information about the OpenCL GPGPU benchmark from SiSoftware, visit the SiSoftware Sandra 2010 website.
Tips and Tricks: Porting CUDA Applications to OpenCLT
If you have some legacy code written in the proprietary implementation, CUDA, and are eager to free your code and development, this article has the tips and tricks you need to start refreshing your code to take advantage of the GPU compute industry standard, OpenCL. This will enable you to not only make your applications portable across vendors, but it also allows you to take advantage of multi-core x86 CPUs in addition to your GPUs. There is also a bit of GPU compute history and OpenCL overview at the beginning of the article that you will not want to miss. View the article.
Coming Soon! OpenCLT Technical Overview Video Series
In the next few weeks, we will post a series of OpenCL technical overview videos featuring Justin Hensley, a senior member of technical staff from the office of our CTO (Advanced Technology Initiatives). In this series, Justin takes developers through an introduction to OpenCL, its design goals and programming model. Later on in the series, he walks you through how to code in OpenCL and touches on important features of the language. This is sure to be an important resource for any developer trying to learn how to code in OpenCL. Stay tuned and be sure not to miss it!
|
|
|
 |
|
Coming in December! CAPS to Release AMD CAL/IL Backend for HMPP
At SC09 this year in Portland, Oregon, CAPS demonstrated the preview of their HMPP development platform with an AMD CAL/IL backend. This extends the existing support the HMPP already has for AMD graphics products. HMPP allows developers to write their compute kernels in Fortran or C with the tool automatically generating the necessary CAL runtime code and IL kernels, reducing the amount of time it takes developers to begin taking advantage of acceleration with ATI Stream technology. Stay tuned for more information on the official release date from CAPS! |
|
|
Available Now! ATI Stream Development Platforms from Colfax and Exxact
ATI Stream developers can now acquire fully configured ATI Stream development platforms from Colfax and Exxact. These new systems provide developers and researchers with complete development platforms, leveraging the latest high-performance AMD GPU architecture along with AMD's OpenCL 1.0-compliant ATI Stream SDK v2.0 to deliver up to 2.72 TFLOPS of compute performance. Visit Colfax's site and Exxact's site for information on their development systems. |
|
|
 |
|
Developer Training Program: OpenCLT Course from VizExperts
VizExperts' OpenCL developer training program is aimed at getting you jump-started with OpenCL. The program includes introductory modules on parallel and GPU computing concepts as well as advanced modules on architecture-focused performance testing and tuning. more information. |
|
|
