Interesting GDC 2005 topics :).

[Panajev2001a]

#1

Introducing the CELL Processor
Speakers: Mark DeLoura (Manager of Developer Relations, Sony Computer Entertainment), Dominic Mallinson (Director, SCE US R&D Sony Computer Entertainment America)
Time/Date: TBD
Track: Programming
Format: 60-minute Lecture
Experience Level: Intermediate - Basic familiarity or some experience.

Description: This presentation gives an overview of the architecture of a CELL Processor, which is a multi-core chip consisting of a 64-bit Power architecture processor, multiple streaming processors, a flexible IO interface, and a memory interface controller. Additional detail is given on the design of a 4-way SIMD streaming data processor in which software controls data movement and instruction flow in order to improve data bandwidth and pipeline utilization.

The quote speaks for itslef.

#2

Advanced Stencil Shadow and Penumbral Wedge Rendering
Speaker: Eric Lengyel (Senior Programmer, Naughty Dog)
Time/Date: TBD
Track: Programming
Format: 60-minute Lecture
Experience Level: Advanced - Requires experience and familiarity with subject.

Description: The stencil shadow algorithm has established itself as one of the mainstream shadow rendering techniques, and it robustly generates very accurate geometry-derived shadows. However, the resulting hard-edged shadow boundaries are becoming less acceptable as hardware rendering power increases and global lighting becomes more realistic. This session reviews advanced implementation techniques of the stencil shadow algorithm and focuses on the relatively new method of penumbral wedge rendering used to generate soft-edged shadows. Light sources in games have conventionally been treated as infinitesimally small points, but the penumbral wedge algorithm can be used to create accurate soft shadows cast by lights having finite areas. The penumbral wedge algorithm is an extension of the stencil shadow technique, and heavily optimized implementations have recently been developed. This session presents all of the relevant mathematics and pixel shader details necessary for soft shadow rendering in real time.

Idea Takeaway: Attendees learn the advanced mathematical and programming details of dynamic soft shadow rendering using optimized implementations of the penumbral wedge algorithm. Example code presented during the session can be directly applied to existing stencil shadow volume engines, extending their rendering capabilities so that they can model area light sources.

Intended Audience and Prerequisites: This session is intended for graphics engine programmers who already have some experience implementing dynamic shadows. Some prior knowledge of the stencil shadow algorithm is helpful in understanding the bulk of the session, even though this topic is reviewed at the beginning of the talk.

Interesting, robust and efficient soft Stencil Shadowing helped by Pixel Shaders... a Senior Naughty Dog programmer... uhm... guess which platform he does have in mind .



#3

COLLADA: An Open Interchange Format for the Interactive 3D Industry
Speakers: Rémi Arnaud (Graphics Architect, Sony Computer Entertainment), Mark Barnes (COLLADA Project Lead Sony Computer Entertainment America)
Time/Date: Thursday (March 10, 2005) 12:00pm — 1:00pm
Track: Programming
Format: Sponsored Session
Experience Level: All - Open to All Levels


Description: *COLLADA* is a *COLLA*borative *D*esign *A*ctivity for establishing an interchange format for interactive 3D applications. COLLADA is co-designed by the major DCC companies such as Alias , Discreet and Softimage , as well as many other middleware companies and game developers such as Emdigo , Ageia , Secret Level , and SCEE . COLLADA specification is defined using XML Schema, and its open-source SDK is available freely at http://www.collada.org. It contains a cross-platform viewer in source code, and several sample code. Each partner writes and supports their own implementation of COLLADA importer and exporter. Source for the latest 3DMax, Maya, XSI are available for download as well COLLADA 1.0 was presented at SigGraph, providing with all the basics. COLLADA 1.1 was released on December 3rd, and contains all the feature requested by game developers such as animations, user-data, and skinning. This GDC'05 presentation will demonstrate usage of COLLADA for the PSP platform and introduction to advanced features such as Physics and Asset management that will be available in COLLADA 2.0.

I expect big improovements in the COLLADA specs with version 2.0 and successive revisions of it.

#4

Effective Use of OpenMP in Games
Speaker: Pete Isensee (Lead Software Design Engineer, Microsoft Xbox)
Time/Date: TBD
Track: Programming
Format: 60-minute Lecture
Experience Level: Beginner - No special knowledge or experience.

Description: Harnessing the power of multiple processors is a new challenge for game programmers, and one that’s likely to be more and more important as the number of multi-CPU gaming systems grows. One technology that allows developers to use multiple CPUs without all the hassle of spawning threads and synchronizing data is OpenMP. OpenMP is a portable, industry-standard API for C/C++ that supports parallel programming. Although OpenMP has been primarily targeted at high-end supercomputers in the past, it’s now available in the Intel compiler, gcc, and Visual Studio 2005. This talk gives a brief introduction to OpenMP, and then dives into ways that games can really take advantage of this technology to improve performance. We cover common errors, portability issues, and debugging.

Idea Takeaway: In order to get the performance they crave, more and more gamers are buying gaming machines with hyperthreading or multiple CPUs. Future game consoles are likely to have multiple processors as well – harnessing the full power of this silicon is a new challenge for game programmers. OpenMP is a flexible, portable technique for programmers to have in their tool chest. OpenMP allows developers to really take advantage of the gaming systems of the future.

Intended Audience and Prerequisites: This presentation is designed for C++ programmers writing games for any multi-threaded platform. The talk assumes basic knowledge of multi-threading programming issues like synchronization and deadlocks, but does not assume the programmer has used OpenMP.

OpenMP, Xbox guy... Xbox 2 tech talks are coming .

#5

Physics in Parallel: Simulation on 7th Generation Hardware
Speaker: David Wu (Director of Technology, Pseudo Intearactive Inc)
Time/Date: TBD
Track: Programming
Format: 60-minute Lecture
Experience Level: Advanced - Requires experience and familiarity with subject.
Description: With the next generation of consoles rapidly approaching, many developers are scrambling to re-architect engines to maximize next generation hardware in an effort to meet next generation expectations. From an architectural standpoint, the most significant change for this generation is Symmetric Multi-processor (SMP) systems. Physics is a computationally intense technology that is becoming increasingly important to games, however, performance issues have always limited it's scope and effectiveness. Fortunately Physical simulation is a technology that is well suited for Parallel implementation. In this lecture, a physics framework designed for SMP systems will be presented, describing implementation issues and trade-offs, performance considerations, and results. Following the presentation a case study will be shown, demonstrating an implementation of the framework within the context of a game that is in development.

Idea Takeaway: The key take away is a broad sample of ideas to help plan physics technology in the next generation of consoles. In particular, the audience will discuss new feature ideas and techniques for implementation.

Intended Audience and Prerequisites: The intended audience is programmers interested in physics and designers with a high tolerance for tech jargon.

Better get used to parallel processing .


#6

So You Need a Better Tools/Technology Group?
Speaker: Ned Lerner (Director of Tools & Technology, Sony Computer Entertainment America)
Time/Date: TBD
Track: Programming
Format: Roundtable
Experience Level: All - Open to All Levels

Description: As games approach animated feature films in quality and cost, tools are becoming a critical part of our business. Yet failure in tools/technology groups is the rule, not the exception. Why and how do we fix ours? What will the next generation platforms demand from our tools/technology group? What has worked and what hasn’t? How are vastly increased content and team sizes affecting the process of game development? How do central technology groups fit into this more movie-like future?

Idea Takeaway: Attendees take away the concept that far better tools are possible. Great tools are expensive, but, if you've got the time, will be worth every dollar they cost. There are many surefire ways of ensuring failure. Some are easy to avoid. Others require daily vigilance.

Intended Audience and Prerequisites: Game developers who need better tools, and wonder why it's so hard to make them.

Tools... Tools... Tools... I hope SCEA shows furing the next-generation that their technology group has improoved and is able to help developers even better.

 

[cthellis42]

Interesting, robust and efficient soft Stencil Shadowing helped by Pixel Shaders... a Senior Naughty Dog programmer... uhm... guess which platform he does have in mind

Definitely the 3DO. Stop being silly!

 

[Fafalada]

Interesting, robust and efficient soft Stencil Shadowing helped by Pixel Shader

Hmm, last I checked penumbra wedges algorithm, mixing it with terms such as "robust" and "efficient" would have made a nice oxymoron.

And it's not like shadow volumes are particularly efficient to start with, soft or not soft.

 

[Ty]

Tools... Tools... Tools... I hope SCEA shows furing the next-generation that their technology group has improoved and is able to help developers even better.

A long time ago I mentioned that Rubin is working on something "cool". This also includes more than a rasterizer.

 

[GameCat]

I don't think you should make any conclusions from the fact that Eric Lengyel is talking about stencil shadows. I had no idea he was at Naughty Dog actually, he has his own software company, Terathon Software that sells a PC-based OpenGL game engine called C4. He's also written in the OpenGL extensions guide and used to be pretty active at OpenGL.org. I had him pegged as a PC guy to be honest, how long had he been at Naughty Dog?

 

[Inane_Dork]

Interesting, robust and efficient soft Stencil Shadowing helped by Pixel Shader

Hmm, last I checked penumbra wedges algorithm, mixing it with terms such as "robust" and "efficient" would have made a nice oxymoron.

And it's not like shadow volumes are particularly efficient to start with, soft or not soft.

Exactly what I was thinking. I'm interested to see what the optimizations are.

 

[Megadrive1988]

oh man, cannot wait. GDC just around the corner. dont have to wait until E3 to learn about at least two of the nextgen consoles. this is gonna be THE GDC to end all GDCs. for the next 5-6 years anyway

 

[Wunderchu]

Art Content Creation for the Next Generation Unreal Engine
Speakers: Jerry O’Flaherty (, Epic Games), Pete Hayes ( Epic Games)
Time/Date: Wednesday (March 9, 2005) 12:00pm — 1:00pm
Track: Visual Arts
Format: 60-minute Lecture
Experience Level: Intermediate - Basic familiarity or some experience.
Description: The next generation of 3D engines for consoles and PCs are built around normal mapping and per-pixel lighting, and thus they bring a substantial change to the visual fidelity of game worlds and characters. New rendering pipelines allow for delivery of easily 100X visual complexity and detail over than the current generation – but with this added benefit comes a staggering increase in production costs. Artist efficiency, rather than technical engine limitations, has become the critical factor. In this talk, we’ll start by showing off assets from two of our unannounced, upcoming Unreal Engine 3 titles, to give the audience an idea of how high the bar will be raised with the new round of game engines. Then we’ll discuss the pipeline, scheduling, staffing, and tools used at Epic Games to rise to this new challenge. We’ll focus on ways to make sure the pipeline process stays on target, since longer production means mistakes are that much more expensive. From the artist’s perspective, this change will cause an industry upheaval like the 2D to 3D transition in the 1990s. We’ll explain how Epic’s staff has made the change to this new technology, and provide suggestions for how artists can prepare themselves for this upcoming career transition.

[206] Negotiating Your Game Development Deal in a PlayStation 3 World
Speakers: Jim Charne (Lawyer, Law Offices James I Charne), Dan O'Connell Offner (Member Offner & Anderson), David S Rosenbaum (Attorney Law Offices of David S Rosenbaum)
Time/Date: Tuesday (March 8, 2005) 10:00am — 6:00pm
Track: Business and Legal
Format: Full-Day Tutorial
Experience Level: All - Open to All Levels
Description: Panelists present a hierarchy of deal points to consider when negotiating a contract, based on the importance of these issues to developers and the potential for success in negotiating them.
Common opening positions from publishers are discurred along with negotiating strategies that can help developers preserve rights and interests in their work.
The day begins with an overview of the state of the business from both a developer and publisher perspective.
The panel identifies developer "survival skills" necessary to grow in a world where there are fewer but bigger A-list projects available for independent developers as publishers consolidate, contract, and rely more and more on in-house teams.
The tutorial concludes with a mock negotiation in which a developer and publisher attempt to make a deal for development of a major console game.

Idea Takeaway: Look to specific areas of each game development contract for opportunities to improve the terms from a developer perspective. By concentrating your efforts in these areas, you may be able to see real gains from what is offered in the first draft prepared by the publisher.

Intended Audience and Prerequisites: Anyone whose responsibilities include negotiation of game development terms: developer and publisher business affairs executives, business development executives, lawyers (California MCLE accreditation will be requested), executive producers who must understand development deals, and students. Basic understanding of game development deal structure is recommended but not necessary.

New Simulation Architectures and Algorithms For Games Physics
Speaker: Michael Lin (President & CEO, Reality Matrix Inc.)
Time/Date: Friday (March 11, 2005) 4:00pm — 5:00pm
Track: Programming
Format: 60-minute Lecture
Experience Level: Advanced - Requires experience and familiarity with subject.
Description: In this lecture, we introduce a cutting edge game physics system. It contains a computational solid and fluid simulation system, with a parallel collision-detection algorithm. It takes benefits from next-gen video game platforms such as Xbox 2 and PlayStation 3, using multiple chips. A more advanced approach is designed to simulate solids and fluids more realistically in real-time. In games, this method can show the crashomg of cars or aircrafts, the impact of glass, or the dropping of brittle materials. In the fluid module, it can generate irregular ocean waves and the coastal waves. This module provides not only the wave shape, but also the hydrodynamic force. By integrating the hydrodynamic force with the solids module, realistic motion of seaborne vehicles can be obtained. If a force-feedback device is connected, the game programmer can create amazing water-based vehicle games such as boat racing and wakeboarding.

Idea Takeaway: From this presentation, the audience can see a new approach to simulating real-world physics in their games. As well as this, they take away an entire picture of how to integrate a physics system in their games and how it works.

Intended Audience and Prerequisites: Those who want to create more realistic physics effects in their game. They may need basic calculus and physics knowledge, as well as programming experience.

No Compromise Cross-platform Audio APIs for the Next Generation
Speakers: Garin Hiebert (Software Engineer, Creative Labs, Inc.), Nathan Charles (Audio API Architect Creative Labs, Inc.)
Time/Date: Friday (March 11, 2005) 10:30am — 11:30am
Track: Audio
Format: Sponsored Session
Experience Level: All - Open to All Levels
Description: Next-generation audio engines will have to accommodate a wide variety of console-specific as well as general purpose audio hardware. This session will describe the best APIs for future audio engines, with an eye towards getting the most out of every platform that the audio needs to run on.

Real-life Case Studies in Multi-threading Game Applications for Performance
Speakers: Leigh Davies (Application Engineer, Intel Corporation), Dean Macri (Engineer Intel)
Time/Date: Wednesday (March 9, 2005) 12:00pm — 1:00pm
Track: Programming
Format: Sponsored Session
Experience Level: All - Open to All Levels
Description: With a large number PCs already supporting Intel’s HyperThreading technology, multi-core CPUs just around the corner, and multi-processor next-gen consoles on the horizon, now is the time to start considering multi-threading your next game engine for performance. Approaching such a task isn’t trivial. In this session we’ll look at several real-life cases where multi-threading was used to garner performance gains. We’ll look at what approaches worked and which didn’t.. Finally, we’ll look at tools that can be used to help and give tips on how to approach multi-threading for your game engine.

Simulation and Rendering of Real-Time Breaking Waves in Water Simulation
Speaker: Adam Lake (Sr. Software Engineer, Intel)
Time/Date: TBD
Track: Programming
Format: Poster Session
Experience Level: Intermediate - Basic familiarity or some experience.
Description: This presentation will present research being done in the industry and academia on the rendering of water. First, an overview of known water rendering techniques will be presented as well as implementations used in commercial games. Next, we will focus on the problem of waves breaking, as in a splash or rolling into a shoreline. Mathematics as well as example implementations, optimizations, and analysis will be presented.

Idea Takeaway: The audience should take away a basic understanding of the current state of the art in rendering water and ocean waves. They will also gain an understanding of the state of the art in breaking waves and how threading and SIMD code optimizations can be applied to create a real-time wave breaking simulation.

Intended Audience and Prerequisites: The intended audience is a graphics architect or programmer engaged in the implementation of game engines that take advantage of the features of next generation hardware. The prerequisites would be an understanding of a basic graphics pipeline. I will introduce the concepts of threading with examples as well as SIMD parallel code and discuss the use of these features in next generation consoles and PC architectures in the context of rendering water and breaking waves. Vertex and Pixel shader implementation will also be discussed.

The Economics of a Next-Gen Game
Speaker: Kathy Schoback (Vice President, Content Strategy, Infinium Labs)
Time/Date: Wednesday (March 9, 2005) 4:00pm — 5:00pm
Track: Business and Legal
Format: 60-minute Lecture
Experience Level: Intermediate - Basic familiarity or some experience.
Description: Rumors of $25m development budgets on next-gen console titles may already be a reality as this talk is presented. And that's before marketing cost! The burning question: how can anyone make money in this scenario? In this talk Kathy reviews an actual next-gen title budget, tweaks parameters that affect title profitability, and (hopefully) suggests ways for all industry parties to get through this morass.

 

[passerby]

No Nintendo. :?

 

[Evil_Cloud]

No Nintendo. :?

Last year there was a presentation by Eiji Aonuma about the history of The Legend of Zelda, right? Or am I mixing up things?

 

[wco81]

It's interesting that there are two talks about physics especially using parallel hardware. You would think if some next-gen games are already in development, they are implementing these ideas long before GDC '05.

Also, even the Intel people are talking about programming models applicable to future consoles even though they will be locked out of them.

Why would someone from Infinium be talking about console game economics? What do they do with the console business now?

 

[Wunderchu]

It's interesting that there are two talks about physics especially using parallel hardware. You would think if some next-gen games are already in development, they are implementing these ideas long before GDC '05.

Also, even the Intel people are talking about programming models applicable to future consoles even though they will be locked out of them.

Why would someone from Infinium be talking about console game economics? What do they do with the console business now?

Some ppl are working on finishing games for the PS2/Xbox/GCN gen. of consoles... (not everyone is working on launch titles, and consoles usually have a ~5 year life-span so devs. have planty of time to start on a next gen. console game in the next few years )

Infinium Labs is launching their "Phantomâ„¢ Game Service" this year ( http://www.phantom.net )