Siggraph Rendermonkey Details and more....

Status
Not open for further replies.
From RenderMonkey pdf:

Creating an environment that is language agnostic allowing any high level shading language to be supported via plugins

Seems to me that RenderMonkey is more of a IDE.
So why not use Cg in RenderMonkey ?
 
Joe DeFuria said:
I think the better question is...

You could use the Cg Language....but why?

Isn't Cg supposed to be able to compile to both DX and OpenGL,
thus removing the need to do 2 codepaths if you'd like to support both
of the API's ?
 
Isn't Cg supposed to be able to compile to both DX and OpenGL

Well, Cg is the language - it abstracted from what you are compiling to; its up to you to tell the compiler whether to generate OpenGL or DX assembly code.
 
http://news.moneycentral.msn.com/ti...&Feed=BW&Date=20020723&ID=1799612

Massive and ATI Join Forces at SIGGRAPH 2002 to Demonstrate Real-Time Cinematic Rendering

ATI Technologies Inc. ATYT(TSX:ATY) and Massive today announced that they will be rendering Academy Award(R) winner, "The Lord Of The Rings: The Fellowship Of The Rings," in real time at SIGGRAPH. This demonstration will be taking place in the ATI booth (#13097) and Massive booth (#5112) using ATI's FIRE GL(TM) workstation graphics technology.
 
Bjorn said:
Seems to me that RenderMonkey is more of a IDE.
So why not use Cg in RenderMonkey ?

Because the Cg language syntax is NVIDIA IP? Because ATI can only write their own compiler, but not their own PROFILE for Cg? Need more reasons?
 
Joe DeFuria said:
I think the better question is...

You could use the Cg Language....but why?
You're kidding right?

For a specific effect, you can get away with 2 or 3 or 4 lines of Cg code instead of the using 20 or more lines (as an example).

Now think about an entire game instead of a specific effect.

RenderMonkey and Cg are nothing but good stuff. Anyone that questions either of them are fools, know nothing of programming or are fanboys.
 
For a specific effect, you can get away with 2 or 3 or 4 lines of Cg code instead of the using 20 or more lines (as an example).

To a lesser or greater degree the same applies for any HLSL not just Cg, which I believe is Joes point.
 
I'm pretty sure he meant as opposed to another HLSL, like the DX9 HLSL, or whatever comes out for OpenGL 2.0...or whatever other HLSL languages Rendermonkey supports now or will support.
 
Well, how many HLSL are there right now that can be used on/for consumer video cards to make games or software such as Alias Wavefront?!? No, that's not Joe's point at all...

DX9 HLSL will be similar to Cg.
 
Reverend again two HLSL languages are coming already:

1) DX9 HLSL
2) OGL 2.0 HLSL

It isn't like CG is the only option here and IMO the Plugin style option is the correct way of approaching these effects. Companies have already invested lots of dollars training their staff on Maya etc...now if a plugin can output shader code on the most popular rendering software without writing a complete new language...what is better ??

We don't need 3 HLSL languages
 
Well, how many HLSL are there right now that can be used on/for consumer video cards to make games or software such as Alias Wavefront?!?

Thats one of the points of RenderMonkey - to bring these higher end HLSL to consumer video cards; i.e. ATi already have compilers for RenderMan and Maya (IIRC).

Further to that, DX9 has its own native HLSL as will OpenGL2.
 
If the ARB merge Cg and 3DLabs's proposal into one language, as has been mentioned in this mornings press release as a possibility, then Cg will be relevant as a High Level Shading Language for both OpenGL 2.0 and DX9 .... I doubt that there's much technically preventing the ARB from doing so, or else they wouldn't have set such a working group in motion....
 
I'm sure ATI would block that in a instant...nothing in it for them..they have allready invested alot of effort into Rendermonkey...

Makes no sense why 3Dlabs would do such a thing unless they want to lose before they even start :-?
 
ben6 said:
If the ARB merge Cg and 3DLabs's proposal into one language, as has been mentioned in this mornings press release as a possibility, then Cg will be relevant as a High Level Shading Language for both OpenGL 2.0 and DX9 .... I doubt that there's much technically preventing the ARB from doing so, or else they wouldn't have set such a working group in motion....

Nvidia did not offer Cg to the ARB, while NVIDIA is seeing OpenGL 2.0 as the "3D Labs High Level Shading Language" (see their Siggraph papers..)

At the last OpenGL ARB meeting, 3D Labs asked for support from the other ARB members. They're looking for someone who implements a DX8/9 compiler for the OGL 2.0 HLSL. This only advantage of Cg (API independence) can be achieved with OGL 2.0 too, we do not need Cg for this. If NVIDIA would support the evolving standards instead of trying to push their own proprietary stuff, everything would go smoother and the adoption of HLSL would be accelerated.
 
If the ARB had big problems with Cg then what was the purpose of today's press release regarding the ARB? To wit:


In addition to worldwide developer support, members of the OpenGL(R) Architecture Review Board (ARB) working group gathered to help define the OpenGL shading language voted unanimously in favor of NVIDIA making a full language specification proposal based on the Cg Language Specification. Next month, NVIDIA will open source the NVIDIA Cg Compiler technology under a nonrestrictive, free license, making the Cg compiler code available to university researchers, programmers, console manufacturers, independent hardware vendors, and other interested parties.

"From developers, hardware companies, middleware and tool vendors -- the interest in Cg has been enormous," said Kurt Akeley, creator of OpenGL and a graphics architect at NVIDIA. "We are delighted that the OpenGL ARB shading language working group has unanimously agreed to review our forthcoming proposal for a Cg based high level shading language."

First Cg Implementations Demonstrated in NVIDIA Booth #8110

Having embraced Cg since its introduction, the digital content creation (DCC) developer community has worked quickly to incorporate Cg into their current development plans. Working directly with NVIDIA to speed their implementations, top content creation software vendors including Alias|Wavefront, Discreet and Softimage will be demonstrating their first Cg implementations publicly for the first time in the NVIDIA booth #8110 at SIGGRAPH. The first Cg-based offerings will be available to DCC customers in upcoming versions of Alias|Wavefront Maya(R), Discreet 3ds max(TM), and Softimage SOFTIMAGE|XSI.

"The industry response has certainly validated the need for a higher level shading language," said Dan Vivoli, vice president of marketing at NVIDIA. "After the first Cg announcement, the traffic on our Web site jumped to six times its normal activity level in the first 24 hours, with nearly a half a million requests to download the toolkit in the first month."

In addition to the aforementioned leading developers, a host of other independent software vendors are actively taking advantage of NVIDIA technology to implement the Cg programming language. Digital Immersion Software Corp., an industry leader in photo-realistic product communication, today announced Presenter 3D(TM) version 1.1 for Autodesk Inventor(R) with real-time rendering made possible with Cg. Aimed squarely at simplicity of operation to increase the mechanical computer-aided design engineer's productivity, this new Cg-infused version of Presenter 3D plugs directly into Autodesk Inventor.

Interactive Data Visualization (IDV), a developer of real-time nature tools for film, video and game developers, is also debuting its Cg-enabled SpeedTree product line during SIGGRAPH 2002. By rapidly integrating Cg into their OpenGL graphics engine in the days following Cg availability, IDV was able to deliver compelling special effects, such as realistic swaying trees, adjustable wind, dynamic shadows and running water on a very tight development schedule.
 
It makes no difference exactly what syntax the OpenGL2 language takes - whether its Cg or 3Dlabs or something else, all that matter is whether the ARB can steer it in the direction the ARB wants as a group. If Cg (or a merger between Cg and 3Dlabs) becomes the language that the ARB decides on then it will have been for the good of the whole ARB, which, you would assume, would mean that the ARB can steer the language and it would not be controlled by NVIDIA (at which point I would see a fracture between NVIDIA's Cg WRT to DX and an ARB controlled language).
 
DaveBaumann said:
It makes no difference exactly what syntax the OpenGL2 language takes - whether its Cg or 3Dlabs or something else, all that matter is whether the ARB can steer it in the direction the ARB wants as a group. If Cg (or a merger between Cg and 3Dlabs) becomes the language that the ARB decides on then it will have been for the good of the whole ARB, which, you would assume, would mean that the ARB can steer the language and it would not be controlled by NVIDIA (at which point I would see a fracture between NVIDIA's Cg WRT to DX and an ARB controlled language).

Yes, I fully agree with you. Not the Cg language syntax or whatever technical reason is the problem, the control of NVIDIA is.
 
And if it happens that way that's good for the industry. And if Nvidia is flexible with Cg, it can be good for NVIDIA too and truly make it a standard , instead of political infighting on profiles and plugins :). BTW I wouldn't suggest using Renderman to do games ....
 
Status
Not open for further replies.
Back
Top