AMD and Havok demo OpenCL accelerated physics

Subject: Graphics Cards | March 26, 2009 - 12:15 PM |
Tagged:

One of the more exciting
technologies we have been in continuing discussion about here at PC Perspective is OpenCL (and DirectX Compute) and how it will change the paradigm for heterogeneous computing.  At the Game Developer's Conference this week AMD and Havok teamed up to offer up some short demonstrations of what I believe is the first publicly shown implementation of cross-processor OpenCL.  It is definitely the first that we are excited about if nothing else.

What the duo is showing off are a couple of subsets of Havok ported to and running on the OpenCL platform.  Both Havok Cloth and Havok Destruction are at least far enough along in development to be shown to the public.  By moving these application APIs to OpenCL, Havok is enabling them to run on any OpenCL compliant processor - an AMD Phenom II or AMD Radeon HD 4870, for example.  They were able to switch on the fly from CPU-driven computing of the physics models to GPU-accelerated physics with a simple button press - indicative of what developers will be able to do in software.  Unfortunately, at this time, performance results of such a switch were not shown as AMD simply stated that today's demonstration was a proof-of-concept and development is still heavily ongoing. 



An example of Havok Cloth

The Khronos group officially ratified OpenCL 1.0 recently that with that complete, AMD, NVIDIA, Intel and others are working hard to get their products driver stacks ready for the new compute language.  AMD now has a working driver for its CPUs and GPUs, though not available to the public, and that is what the Havok physics APIs be demonstrated are utilizing.

The power behind this announcement is easy to see - now that Havok can run its APIs easily on either a CPU or GPU, the ability to accelerate physics in a way we had hoped would come about with technology like AGEIA PhysX is much more attractive.  While currently today PhysX only runs on NVIDIA GPUs, OpenCL products will run on AMD and NVIDIA GPUs as well as Intel and AMD processors enabling heterogeneous computing algorithms across both product lines.  This will also make it much likely that developers will take the additional time to program for such accelerated physics as the install base will have essentially doubled. 

As for addressing the "effects physics versus gameplay physics" debate, the OpenCL implementation of Havok's APIs might help with this as well.  The software developer will be able to query the system it is running on to determine how much processing power it actually has (based on predefined standards from an OpenCL "host") and adapt the algorithms accordingly.  If a gamer has a slower CPU but a really fast GPU, for example, the physics models might be able to be increased dramatically; if a user has a higher end CPU but lower end GPU then the same algorithms could be run, just slightly slower, and likely adapted down in quality effects. 

This is still a work in progress from all parties - OpenCL driver stacks from hardware vendors and the first OpenCL software from developers such as Havok and others.  It will be VERY interesting to see how this progresses as we get into 2009.

AMD Demonstrates Optimized Executions

of Havok Middleware on AMD platforms

– Balanced Platform of CPU + GPU
Processing Delivers Optimal Game Experience –

SAN FRANCISCO, Calif.— March 26, 2009— Advanced
Micro Devices,
Inc.(NYSE: AMD) and Havok, the premier provider of
interactive software for physics simulation and content development,
are presenting new,optimized executions of Havok’s physics middleware on AMD
platforms at the 2009 Game Developers Conference.  The demonstrations include the
first OpenCL supported execution of Havok
Cloth™.
 

Havok offers a complete modular
suite of products that help visual and interactive content developers create
more realistic games and cinematic
special effects. As the latest software developer to take advantage of ATI Stream technology to leverage multi-core architectures and accelerate
execution of highly parallel functions, like real-time cloth simulation, Havok
will enable game developers to offer improved performance and interactivity
across a broad range of OpenCL capable PCs. AMD has recently introduced optimized
platform technologies, such as “Dragon
desktop platform technology, which balance performance between the CPU and GPU
with ATI Stream technology to deliver outstanding value.

 “Havok is committed to
delivering highly optimized cross-platform solutions to our  game customers and
we are pleased to be working with AMD to ensure that gamers enjoy a great user
experience when running Havok-powered games on AMD platforms.,” said
David Coghlan, vice president of development for Havok. “Unlocking the parallel processing capability
of AMD’s hardware provides real advantages to our customers, and the
greater the total computing resources available, the better the gaming experience
developers can deliver.”

"Havok’s awesome toolset
has allowed us to deliver astonishing physics interactions in our games,
including detailed real-time destruction and complex ragdoll models, and we are
excited about using ATI Stream technology to pursue more astounding in-game
accomplishments,” said Andrey Iones, chief operating officer of Saber Interactive.
“We are excited that AMD and Havok are working together and leveraging an
open standard like OpenCL.”

Additional detail on ATI
Stream technology can be found at www.amd.com/ati.stream

About
AMD

Advanced
Micro Devices (NYSE: AMD) is an innovative technology company dedicated to
collaborating with customers and partners to ignite the next generation of
computing and graphics solutions at work, home and play. For more information,
visit http://www.amd.com.

No comments posted yet.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

By submitting this form, you accept the Mollom privacy policy.