Review Index:
Feedback

Ray Tracing and Gaming - One Year Later

Author: Daniel Pohl
Subject: Processors
Manufacturer: Intel
Tagged:

Progress

After my last article about ray tracing and gaming hit the web I did some more research on scaling of ray tracing with the number of CPU cores. In order to simulate a 16-core machine I took four quad-core PCs and connected them over a Gigabit-Ethernet to combine their power. Because my project used the ray tracing library OpenRT from Saarland University which supports distributed rendering, this was quite easy to achieve.


Performance scaling of ray tracing with different number of cores

The results were amazing: If you use a 16-core machine instead of a single-core machine then the frame rate increases by a factor of 15.2!

After all the attention the previous ray tracing article got around the world I was contacted by several companies interested in this technology. One of them was Intel. They told me they would have a real-time ray tracer that would be around 10× faster than everything else that has been published so far. These performance numbers were already written down in some research papers, but I did not trust them without seeing it myself. So I went over to Santa Clara to get a live demonstration of it.

 

Then I saw it trace,   

now I'm a believer!


    

So now I am a full-time research scientist working for Intel on ray tracing. As it turns out, Intel’s labs are very interested ray tracing because it is so well suited to general purpose CPUs. Our ray tracing research is just a part of an overall program here called “tera-scale computing” aimed at scaling CPUs from a few cores to many (meaning tens or hundreds).

 

Joining Intel also gave me the chance to demonstrate Quake 4: Ray-Traced with the new Intel ray tracer on several occasions such as the Games Convention 2007 in Leipzig and Intel Developer Forum  Fall 2007 in San Francisco. At the latter event it was even featured in the keynote from Justin Rattner (Intel CTO) about virtual worlds.


At HD resolution we were able to achieve a frame rate of about 90 frames per second on a Dual-X5365 machine, utilizing all 8 cores of that system for rendering.

More details about the presentation at Fall IDF 2007 can be found in Ryan Shrout's article at PC Perspective.

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.