Last day at GDC and some graphics thoughts
 |
| GDC 2008 |
Today is the last day of GDC. In the past two days food quality has slightly improved. Though after the 1st day, the less expensive pass holders started coming: a lot more people !
There are very long lines to get the lunch boxes, but luckily they go pretty smooth.
A combination of jet-lag and night coding sessions brought me to a completely screwed schedule, with the advantage of not having problems waking up in the morning.. because by 4 am I'm already up (whether I slept 10 or 3 hours 8).
Some lectures were pretty interesting. I've been going to multi-core programming lectures (lots of them, sponsored by Intel) rather than more traditional lectures.. and completely avoided most non-technical lectures.
I bought a RenderMan book and I started writing a simple RenderMan interface !
There is a lot of talk about off-line rendering going real-time, but I don't think that OpenGL and Direct3D 10, 11, 12 are the right way to go about it. There are a lot of things that we 3D programmers on the real-time side have to learn from the pre-rendered world if we don't want to be always lagging behind the pre-rendered movies.
Of course pre-rendered will always win in terms of quality, but it's important to embrace the off-line rendering systems build pipeline and then go on about tricks to scale quality down without having to rebuild models, textures and shaders from scratch.
I see RenderMan the way I saw OpenGL 10 years ago. OpenGL was considered overkill for games, until Carmack started using it with the support of 3Dfx.
If Carmack today said we should be using RenderMan, I bet everyone would rush to it.. but it's me the one that is saying it.. 8)
- Davide's blog
- Login to post comments
RenderKaz
I think I had mentioned in a previous post about how there will be some efforts to bring some of the off-line stuff into real-time; it's pretty obvious I believe. You're right about embracing the build pipeline; it's important to build a good foundation.
Because of economic reasons, I think the progress to a Real-time Renderman system (if that should happen) will be through OpenGL and Direct3D . There's just way too much R&D time and money invested to just suddenly abandon the current methods. I believe A Real-Time Procedural Shading System for Programmable Graphics Hardware used OpenGL with multi-pass rendering before GPUs had vertex, pixels, or unified shaders. An interesting point is that they used RenderMan as the shader language. ;-)
RenderMan is possibly the right way to go ... especially since there is likely to be some benefit for the CG movie shops of being able to do some things in real-time. By the way, RenderMan is an of the REYES algorithm. ;-)
I said RenderMan Interface,
I said RenderMan Interface, not RenderMan in general.
REYES is beyond the scope even though it's possibly around the most efficiently system out there.. still it does not pretend to maximize efficiency nor it does.
Note to self: Wear glasses
Note to self: Wear glasses when reading kazzuya.com .... 8P
You girls should look at
You girls should look at OpenRT. I have seen it in action in commercial product before and it was impressive. Also my good friend was one of the core developers so I know a lot of details. Of course back then it was running on farm of Linux machines but if Intel brings this 80 cores CPUs then this OpenRT will just ready solution for real time raytracing.
Master Duddie, I've seen
Master Duddie, I've seen OpenRT, that's why I keep mentioning REYES and potential possibilities with it. Kaz says, "Bah, it won't happen any time soon because of (insert reason)" ... 8P I can't really raise any counter-arguments to what he says since I realize that he may be doing stuff that backs up his opinions, and I am merely using my intuition to attempt to divine where things are going future-wise.
I still feel that real-time ray tracing (sure, maybe not at the Pixar level ... yet) may be possible. I think Daniel Pohl might have done work on OpenRT or SaarCOR, but I can't say for sure.
A pattern that I've noticed about R&D of this nature (as more people start to look for new ideas/technologies) is that someone usually discovers some fundamental technique or idea that makes the technology possible ... or ways to lower the cost of implementing such technology is discovered.
Perhaps what will happen is that (in the near future), you'll have these 80 core CPUs that one does the ray tracing with in software. Then at some point when the technology starts to become well know and costs drop, there may then be hardware ray tracers. Kind of like how one did software 3D engines a several years back, but not a lot of that is done in hardware.
Interactive Ray Tracing in Games sounds pretty interesting. Too bad I didn't have the opportunity to be there in person.
Whatever the results, I think Intel will benefit greatly from whatever the R&D effort produces as it's a good real-world problem that's ideal for multi-core/multi-processor architectures.
Good morning M. Duddie
Good morning Mr. Duddie 8)
Haven't we mentioned this real-time ray-tracing about a 100 times already ?
Real-time ray-tracing's rendering quality is bad and the only real good thing about it is the reflections and refractions and hard shadows..
See this for a better explanation http://www.beyond3d.com/content/articles/94/1
tryinghelp.wav
Maybe... I was not reading carefully enough :) Anyway, for me the most graphically demanding game recently is Solitare so... :P
Where there's a will ...
Hey RenderKaz,
Check out these articles if you have time. They're quite interesting:
Ray Tracing and Gaming - One Year Later
Real Time Ray-Tracing in your Pocket
Now I'm not saying that "Ray Tracing is the future!", but more I find it interesting how much effort Intel seems to be putting behind R&D in this area.
Ummmm.. the only articles
Ummmm.. the only articles floating around are those of Daniel Pohl. Hardly much effort from a company such as Intel. I'm pretty sure that they are not betting everything on Ray Tracing.
I think Intel uses this Ray Tracing thing as a gimmick to impress people on one side and maybe even to divert attention where maybe the real focus is (?)
It's great to promote research in that sense too.. but to me it's only some basic research, hardly a solution to photorealism, not unless those real-time ray tracers start employing some solid global lighting pre-calculations ;)
Master Plan
My impression is that Daniel Pohl's articles are the one mostly seen mainly because he had started writing such articles when he did his real-time ray-tracing work with Quake 4: Raytraced before he got hired by Intel and that him writing articles are just a continuation of what he was already doing. I hope I didn't imply that they are betting everything on ray-tracing.
I think it's interesting that they recently bought the Havok Engine and the Offset Engine:
Intel Acquires Offset Software, Project Offset Engine
Why Intel bought Project Offset and the Offset Engine
Intel Purchases Game Tools Firm Havok
I believe that the ray-tracing, Offset Engine, and Havok are applications that align well with their multi-core efforts (short term and long term). I mean not too long ago nVidia bought AGEIA. It just seems that Intel has some plans to push some kind of technology (I believe it's their Tera-scale Computing effort) in a very serious and tangible way.
I guess we'll have to wait about 1-2 more years and see what they're actually up to. 8P
I dont think they are up to
I dont think they are up to anything really. Just a big company doing research with extensive funding.
Intel is definitely on the
Intel is definitely on the move, but I wouldn't associate Project Offset with Havok.
Havok is a proven physics library, it works well because physics in games doesn't change as much as graphics does. Project Offset however is more an hyped game with some cool shadows, HDR and so on.
They are on the move trying
They are on the move trying to get a share of the CG market, but it will not make or break them. I assume their main revenue is still from sales of the pentium processor.
RenderKaz, I mentioned Havok
RenderKaz, I mentioned Havok along with Offset because they were among their recent purchases. Obviously Havok has nothing to do with graphics, but it looks like it can definitely be used to take advantage of their Tera-Scale efforts ... in which the ray-tracing R&D stuff falls under. It seems that the goal isn't graphics per se, but finding potential commercial applications for that multi-processor technology. At the moment though, I'm a bit intrigued by Larrabee ... though really saddened that it's x86 based ... but if one has already proven technology laying around, there's very little point in spending extra money on making something new ... I suppose this can be a good or bad thing ... 8P
On a different note, iPhone development sounds pretty interesting: iPhone SDK Available Today for Free, $99 to Publish Your Apps
Larrabee
Was doing some more poking around and I found these links:
Meet Larrabee, Intel's answer to a GPU
Clearing up the confusion over Intel's Larrabee
It will be interesting to see what happens in the next few years. :)