Layers of abstraction always exist between man and machine. Starting with our gestures and speech, one symbolic language passes through an interface to another, ultimately translated to code.
The fundamental language or instruction set of an x86 central processor evolves slowly, balancing new features and compatibility. Today, applications are generally written and compiled directly for the CPUs low level instructions.
ATI Stream technology from AMD is a set of AMD technologies that allow the hundreds of parallel Stream cores to accelerate general purpose applications. It embraces open, standards-based approaches to accelerated, highly parallel processing on ATI RadeonTM HD general purpose GPUs. This technology is well suited to operations performed on massive data sets undergoing rapid and consistent processing. The applications have the potential to transform how we use computers.
Unlike an x86 processor, a GPU employs a software interface allowing it to evolve rapidly while retaining compatibility. Improvements in performance can be delivered by software drivers that take better advantage of the underlying hardware, and allow applications to have a predictable interface to such hardware. With the addition of standards-based interfaces, a GPU is now able to serve the dual purpose of specialized graphics processing and accelerated computation.
AMD is developing general purpose GPUs and the software technology that translates these standardized interfaces into fast, efficient, parallel code. Innovation in the underlying silicon and low level instruction sets can continue while maintaining code stability through an abstraction layer. AMD supports both Microsoft’s DirectX 11 featuring DirectCompute and now OpenCL.
Developed in an open standards committee with representatives from major industry vendors, OpenCL gives users what they have been demanding: a cross-vendor, non-proprietary solution for accelerating their applications on their CPU and GPU cores.
ATI Stream technology can excel in applications where the underlying data is a representation of physical reality. An abundance of visual information. The interaction of forces of nature. Physical representations in space and time. With ATI Stream, OpenCL, and DirectCompute we now have new tools to face the ultimate challenge, the one that useful and “intelligent” machines must learn to face – how to deal with reality.
For more on the introduction of the ATI Stream SDK 2.0 with OpenCL support check out Nigel Dessau’s new blog.
If you are interested in learning how OpenCL works, read on for Part 2, where Ben Sander and I discuss the power of ATI Stream technology and the elegant, standards-based interface now available with OpenCL for GPU. We explore the relationship between, OpenCL, AMD’s traditional support for industry standards, and the new ATI Stream SDK v 2.0.
Read On -> Dealing With Reality | The Introduction | ATI Stream Technology and OpenCL | Part 2
Be sure to check out Simon’s ongoing series on The Digital Nexus.
Simon Solotko is a Senior Advanced Marketing Manager at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.


(11 votes, average: 4.55 out of 5)
#1 by asH - October 21st, 2009 at 04:17
I think most folk can’t see the potential paradigm shift in the direction AMD is moving, all the while remaining consistent to their ecco-beliefs- “less power more production”…Scotty?. Their Opteron multicore processor stresses low power with multiple cores- I think it’s clear, looking at the Opteron and their other recent chips, that AMD is not competing with Intel for the top speed honors, or the fastest Giga chip. We also see perhaps closer working relationships between AMD/ Apple, AMD /Microsoft, connected by AMDs graphic card products 4000+ series, OpenCL, directX11,and DirectCompute… THE GIGA-CHIP WARS ARE OVER! AMD has been preparing for this end-of-day; future is now scenario for a while now. IT will no longer be about racing to pump up the power of the CPU (MORE PORER SCOTTY!), because all the processing power one ever needs goes unused, untapped. In each of our computers has a reservoir of processing power we’ve yet to unleash; we have had it for some time now- it’s in the GPU. Coupling the GPU to current low watt multicore processors, through OpenCl, DirectX11, or direct-compute, CPUGPU becomes realized as both redeemer and the dragon slayer simultaneously. The second stage is the physical bonding between GPU & CPU- the all in one Llano
On another note- I personally don’t believe Larrabee will be developed in time, if at all work as promised. I believe they were late in seeing, and are desperately shoveling recourses to catch-up ..but this is just me
asH