bit-tech.net

New graphics firm promises real-time ray tracing

New graphics firm promises real-time ray tracing

Caustic says the CausticOne will offload ray tracing calculations from the CPU or GPU and then feed back the data, enabling your PC to shade a ray traced scene in up to 20x the speed.

The potential for real-time ray tracing has been a major talking point in 3D graphics ever since Intel announced its plans for the x86-based multi-core Larrabee discrete graphics chip at IDF in 2007. However, a new graphics company called Caustic Graphics reckons that it’s already uncovered the secret of real-time ray tracing with a chip that it says “enables your CPU/GPU to shade with rasterisation-like efficiency.”

Caustic’s management team includes previous employees of Autodesk, Apple, ATI, Intel and Nvidia, and the company has already developed a chip for accelerating ray tracing. Called the CausticOne, the new chip isn’t targeted at gamers, but is instead aimed at 3D professionals in the film, game development, automotive and consumer product design industries. This isn’t surprising, given that current gaming APIs are based on rasterisation, but Caustic is still making some impressive claims about its technology, while not being afraid to rubbish the efforts of other graphics companies when it comes to ray tracing.

“Some technology vendors claim to have solved the accelerated ray tracing problem by using traditional algorithms along with GPU hardware,” says Caustic, referring to companies such as Nvidia which recently demonstrated real-time ray tracing using CUDA . However, the company adds that “if you've ever seen them demo their solutions you'll notice that while results may be fast—the image quality is underwhelming, far below the quality that ray tracing is known for.”

According to Caustic, this is because the advanced shading and lighting effects usually seen in ray-traced scenes, such as caustics and refraction, can’t be accelerated on a standard GPU because of incoherent rays. Caustic explains that “primary rays will bounce and scatter secondary rays throughout complex scenes. These secondary rays may be incoherent (they reflect off in various directions) and it is these incoherent rays that are the key to creating the advanced effects required to achieve photorealism in complex images (numerous objects, refractive and reflective materials, rough surfaces).”

Conversely, Caustic claims that the CausticOne “thrives in incoherent ray tracing situations: encouraging the use of multiple secondary rays per pixel. Its level of performance is not affected by the degree of incoherence.” Explaining why traditional algorithms and parallel processing isn’t the best system for ray tracing, Caustic said that they “ignore memory bandwidth, cache performance and parallel compute utilisation. Only primary rays get a performance benefit with this approach.”

To accompany the CausticOne, the company is also introducing its own API, called CausticGL, which is based on OpenGL/GLSL. The API will feature Caustic’s unique ray tracing extensions, and Caustic says it will also “enable developers to developers to create industry-specific rendering solutions via their own existing renderer.” The company says that the CausticOne (which has 15 patents pending on the technology and algorithms used) basically off-loads raytracing calculations and then sends the data to your GPU and CPU, enabling your PC to shade a ray traced scene much more quickly.

The CausticOne is due to be released in April 2009, and Caustic claims that it will speed up the ray tracing process by an average of 20x. However, the company is making wilder claims about its second-generation Caustic product, due to be released early next year, which it says will offer “200x the speed over today's state-of-the-art graphics products.” According to Caustic, the huge speed jump is due to the company implementing its own patent-pending ray tracing algorithms in a piece of silicon.

Explaining the need for ray tracing acceleration, Caustic Graphics’ CEO, Ken Daniels said that "for years, 3D professionals in multiple industries have laboured under the yoke of slow iterations and unwieldy offline render farms." According to Daniels, "Caustic puts the power of a render farm, operating at interactive speeds, on every desktop, enabling designers and animators to get from concept to product faster, better and at lower cost."

Meanwhile, Jon Peddie of Jon Peddie Research was keen to congratulate Caustic Graphics on its new technology. "Demos have been done with 16 or more processors, super computers, and other esoteric devices,” said Peddie, “but never anything that was within reach of a PC budget. Caustic Graphics has made the breakthrough with a combination of a small hardware accelerator and some very innovative software to be able to deliver real-time, complex, high-resolution ray traced images - this is an amazing accomplishment."

Do you work with ray tracers in your field of work, and would you be interested in a ray tracing accelerator such as this? Plus, when do you think we’ll ever see real-time ray tracing in games? As always, let us know your thoughts in the forums.

25 Comments

Discuss in the forums Reply
GFC 12th March 2009, 14:02 Quote
It's for pros only.. not really anything interesting for the average guy.. :/
perplekks45 12th March 2009, 14:11 Quote
The next PhysX?
p3n 12th March 2009, 14:17 Quote
Seems odd to make a product that would surely come close to saturating whatever BUS it plans to communicate on, another chip best sat next to the CPU die imo (gogo intel/amd)!
Star*Dagger 12th March 2009, 14:18 Quote
I am interested. I'd like to see raytrace level games and OS's.
Bauul 12th March 2009, 14:25 Quote
Pictures or not true. Whilst in theory it sounds great, this kind of "miricale breakthrough" by a tiny, unheard of company rarely results in anything worth talking about. Remember the tiny start-up that invented perpetual motion?
Skiddywinks 12th March 2009, 14:28 Quote
Quote:
Originally Posted by perplekks45
The next PhysX?

No. Caustic are using an open source API here. Anyone and their dog can have a go, which is great for the market and why PhysX failed.
DarkLord7854 12th March 2009, 14:30 Quote
I bet it'll get snapped up by a bigger company.. probably Intel since they've invested so much into raytracing
Yemerich 12th March 2009, 14:36 Quote
Optical physics have always benn the great vilain for people that worl with rendering. You can jump from a 5 min rendering scene to a 15 min scene just by adding some secondary bounces. And indeed it makes a HUGE difference. PhysX is "just" an algorithym for "mechanical" phisics, we are talking here of adults play.
I really doubt this kind of stuff will hit games anytime soon (at least 2 years on the go), because the actual engines would have to adapt TONS of codes to include this new algorithym. But again, its just matter of time untill we see games rendering real-time with the quality of a "Beowulf, the movie"-like game.. Perhaps a LONG time, but still just a matter of time
HourBeforeDawn 12th March 2009, 17:54 Quote
this is what nVidia is really afraid of and why they keep bad mouthing ray tracing as it uses x86 technologies which Intel and AMD/ATI wont have issues implementing into their graphics cards and why nVidia is trying to make their own x86 cpu dispite bad mouthing how "useless" it is lol but thats nVidia for you, bad mouth what you cant do to trick the people but secretly try to make it because you know thats where things are heading.
[USRF]Obiwan 12th March 2009, 20:01 Quote
Well like everything else I stand by: "seeing is believing"
eeevan 12th March 2009, 20:11 Quote
Quote:
Originally Posted by Bauul
Pictures or not true. Whilst in theory it sounds great, this kind of "miricale breakthrough" by a tiny, unheard of company rarely results in anything worth talking about. Remember the tiny start-up that invented perpetual motion?

Oh come on... What a pathetically laughable statement.

Many a "tiny, unheard of company" have evolved into massive corporations after breaking into the market with a catalystic, breakthrough technology.

The fact that you cited a perpetual motion related startup just makes me lol.


fail.
SuperNova 13th March 2009, 08:13 Quote
Well its always nice with new inventions, not to mention another competitor on the market. Even though i have some questions about how fast this card actually will be. They say 200 times today's fastest graphiccard which are highly parallel. If we take a single card, (not X2, GX2 and so on) they have about 1Tflops (4870 @ 1.2). I'm having a hard time to see a card with the equivalent of 200Tflop coming out ext year... Therefor it would be really fun to know the estimated computation power of this card, just to see how much comes from optimizations in the soft and hardware.


With a card specially designed for raytracing will probably not handle rasterization very well (if the cards don't have 200Tflop of raw power of course and emulates, like larrabee is said to do). This will probably be the the biggest problems for entering the gaming industry since it will limit the use to games relying solely on raytracing. For the 3D-modeling industry it will probably do wonders (if it performs as promised).

5 computers with 4 cards each would give the equivalent of 4 Pflops (double that if they have 4870x2 in mind when saying 200x), I7 965 is around 50Gflops.

But no mater ow they achieve this speed its great!
perplekks45 13th March 2009, 10:38 Quote
They won't enter the gaming market I guess, not before 2010. And even then it'll be hard for them.
Bauul 13th March 2009, 14:42 Quote
Quote:
Originally Posted by eeevan
Oh come on... What a pathetically laughable statement.

Many a "tiny, unheard of company" have evolved into massive corporations after breaking into the market with a catalystic, breakthrough technology.

The fact that you cited a perpetual motion related startup just makes me lol.


fail.


Aww bless. Come on then, give me an example of a tiny start-up company that promised an incredible technological breakthrough none of the market leaders had managed, that actually turned out to a) work, b) be commercially viable and c) not a complete failure.

Either they're exagerating the benefits of the new software, or they'll be snapped up by one of the big players faster than you can say "ray-tracing", and we'll never hear of them again.
steveo_mcg 13th March 2009, 14:52 Quote
Quote:
Originally Posted by Bauul
Aww bless. Come on then, give me an example of a tiny start-up company that promised an incredible technological breakthrough none of the market leaders had managed, that actually turned out to a) work, b) be commercially viable and c) not a complete failure.

Either they're exagerating the benefits of the new software, or they'll be snapped up by one of the big players faster than you can say "ray-tracing", and we'll never hear of them again.

Google, MS err nah thats the only two i can think off

Though both aided by being an entry in to a fairly new market.
gmmail1980 13th March 2009, 15:42 Quote
DXR_13KE 13th March 2009, 16:56 Quote
i can expand its memory?
Peet42 14th March 2009, 00:07 Quote
Quote:
Originally Posted by Bauul
Remember the tiny start-up that invented perpetual motion?

What? Aren't they still going...? ;-D
Anakha 14th March 2009, 00:33 Quote
They're going to have major problems with their pending patents. Look at Saarcor, the RPU, and OpenRT.
Elton 14th March 2009, 05:02 Quote
Now if friggin Lucid would release their Hydra Chip in the P55 or the X58, it'd work out great.
roshan 14th March 2009, 14:06 Quote
what happened to to original intell labree ray tracer graphic chip.?
we cant believe what they say until we see the demo of it ray tracing.
perplekks45 14th March 2009, 14:12 Quote
Larrabee is still in development and will be released... whenever Intel wants to release it. Probably later this year.

And yes, I'm looking forward to it. Personally I don't need no ray tracing card but there's a market for it. I hope BT gets an example to test it. ;)
Dreaming 15th March 2009, 13:40 Quote
Quote:
Originally Posted by Skiddywinks
No. Caustic are using an open source API here. Anyone and their dog can have a go, which is great for the market and why PhysX failed.

I thought Nvidia bought it up?
simond 15th March 2009, 23:45 Quote
Quote:
Originally Posted by Bauul
Pictures or not true. Whilst in theory it sounds great, this kind of "miricale breakthrough" by a tiny, unheard of company rarely results in anything worth talking about. Remember the tiny start-up that invented perpetual motion?

I remember that company - I think they're still going............
Rumble 16th March 2009, 15:46 Quote
Quote:
Originally Posted by Skiddywinks
No. Caustic are using an open source API here. Anyone and their dog can have a go, which is great for the market and why PhysX failed.


They failed so badly that Nvidia just had to acquire them?

:?
Log in

You are not logged in, please login with your forum account below. If you don't already have an account please register to start contributing.



Discuss in the forums