|
Intel Game Threading Tutorial (All-day session)
TUESDAY, MARCH 24
Take part in the latest installation of the popular Intel tutorials! This year, we’re back to a hands-on format where participants will review threading architectures in games to practice “thinking with threads”, build a new system for an existing threaded game framework, experiment with data dependencies between game systems running in different threads and measure bottlenecks and fix them. Presented by Orion Granatir, Paul Lindberg and Brad Werth.
- Download the session files (190MB .ZIP)
- Download the session PDF materials (7.18MB .ZIP)
- Watch the Take Five videos of the session, starting with the Introduction
First Peek: New Intel Graphics Performance Analyzers
WEDNESDAY, MARCH 25
Discover the new graphics tools just released by Intel and realize the full performance potential for your game title! Developers will find out where these tools came from, why they are here, where they are going in the future, and how it’s all going to change everything. Presented by Aaron Davies, Cian Montgomery, and David Shinsel.
- Download the session materials
Programming Tips for Scalable Graphics Performance
WEDNESDAY, MARCH 25
This course will focus on programming for scalable graphics applications. It will cover performance considerations when programming for Integrated Graphics in general with specific tips for Intel® Integrated graphics. The class will include code examples for shader and hardware optimization with performance measurements. Presented by Luis A. Gimenez and Ganesh Kumar.
Kaboom: Real-Time Multi-Threaded Fluid Simulation for Games
WEDNESDAY, MARCH 25
Learn how to easily add real-time 3D smoke, fog and other fluid simulations to your game without using up the GPU. In this session, we will present the source code to a fluid simulator optimized for multi-core CPUs. Our simulation source code can easily be integrated by game developers into their engines to produce unique 3D effects. Traditional methods of fire and smoke rendering typically involve billboarded particle systems, imposters, and other 2D-based viewpoint simulations. While these methods produce acceptable visuals, they are 2D imposters in a 3D world. Modern multi-core processors are now powerful enough to deliver compelling 3D representations of gases and liquids. The Navier-Stokes equations produce a beautiful and naturally evolving simulation and have been used in offline rendering for movies, engineering applications, and research. With some simplification coupled with the increasing power of modern, multi-core processors, the application of the Navier-Stokes equations for fluid simulations in real-time is now possible. Attendees will learn how to integrate an efficient fluid simulation that scales with modern multi-core processors into their code, providing real-time fluid simulations tailored to model smoke, fire, and fog. Presented by Jeff Freeman.
- Download the session materials
Threaded AI For the Win!
WEDNESDAY, MARCH 25
Writing threaded AI is epic? Threading AI is easier than you think! We will examine how AI can be threaded and live in a highly parallel environment. How can you thread AI? How can AI talk to physics running on another thread or device? Is deferred processing worthwhile? Do you have to thread your designers? This presentation will answer these questions and more. The presentation will include a quick overview of our Intel Smoke demo; Smoke is a n-way threaded framework that includes source code for highly parallel AI. Presented by Orion Granatir.
Accelerating Creativity: Building Environmentally Aware Characters with Havok Behavior
WEDNESDAY, MARCH 25 and THURSDAY, MARCH 26
Get an in-depth look at Havok's latest Havok Behavior product release, accelerating creation of sophisticated behaviors that combine animation, physics, environmental sensing & constraints, inverse-kinematics, paired-animations, and real-time clothing preview using Havok Cloth. This talk includes a presentation of a new Havok product launching this GDC. Presented by Jeff Yates.
Havok Technologies: Present and Future
WEDNESDAY, MARCH 25 and THURSDAY, MARCH 26
We present a detailed look at Havok's runtime and tools technologies including Physics, Destruction, Cloth and Character tools. We will showcase all our latest product and technology advances, and will outline our plans for 2009 and beyond. This talk includes a presentation of a new Havok product launching this GDC. Presented by Andrew Bond.
Optimizing Game Architectures with Intel Threading Building Blocks
THURSDAY, MARCH 26
This talk will show how the cross-platform threading library, Intel Threading Building Blocks, can enhance the performance of your game. Techniques will be shown starting from some simple work-alike replacements up to a full, transparent replumbing of your threaded architecture. You'll see how TBB can impersonate the most popular threading patterns while delivering enhanced performance on PC, XBox360, and OS X. Presented by Brad Werth.
Download Optimizing Game Architectures with TBB
Download the session session code samples
Taming Your Game Production Demons: The Offset Approach
THURSDAY, MARCH 26
Time and cost are the demons that plague every game. We have created a content-centered development team, focusing on tools that magnify the impact of each individual artist and remove barriers to creativity. Offset’s editor and engine, still in production, are being designed and developed from the ground up to enable fast iteration with a WYSIWYG interface for optimal artist efficiency. In this presentation we will share the steps (and some missteps!) we took to give our artists the tools they need to tame the demons of their production pipeline. Presented by Remi Arnaud and Ian Lewis.
- Download the session materials
Who Moved the Goalposts? The Rapidly Changing World of CPUs and Optimization
THURSDAY, MARCH 26
In less time than some games take to develop, Intel desktop PC architecture has undergone two major revisions. This talk provides an insight into some of the major changes and the impact on program optimization. Innovations that drastically affect PC optimization techniques include the reintroduction of Hyper-threading, a new cache design, a new memory controller with significantly enhanced bandwidth and several new additions to the SIMD instruction set. As well as looking at the physical changes and the reasons for them, the session will also look at how to detect the architectural changes and typical optimizations and how to mitigate the problem of divergent performance paths. The session will finish with a look at the future and some of the technologies that could be reality before games just starting out today come to market. Presented by Leigh Davies.
Rasterization on Larrabee: A First Look at the Larrabee New Instructions (LRBni) in Action
FRIDAY, MARCH 27, 9AM; Moscone North Hall, Room 135
Larrabee is Intel's revolutionary approach to take the current evolving programmability of the GPGPU to its logical end. The Larrabee architecture features many cores and threads, as well as a new vector instruction-set extension, the Larrabee new instructions (LRBni). This talk will provide an overview of LRBni and discusses the major instruction features - 16-wide SIMD, multiply-add, ternary instructions, predication, built-in data-format conversion, and gather/scatter. The talk will then take a close look at a specific - and not obviously vectorizable - application of LRBni - rasterization. This is a crucial stage in the Larrabee rendering pipeline, and it demonstrates how developers can use the flexibility of the new instruction set to solve problems that are not obviously "shader-like". Presented by Michael Abrash.
SIMD Programming on Larrabee: A Second Look at the Larrabee New Instructions (LRBni) in Action
FRIDAY, MARCH 27, 10:30AM; Moscone West Hall, Room 3002
Larrabee is Intel's revolutionary approach to take the current evolving programmability of the GPGPU to its logical end. The Larrabee architecture features many cores and threads, as well as a new vector instruction-set extension, the Larrabee new instructions (LRBni). This talk follows Michael Abrash's first glimpse into LRBni and examines the programming methods and hardware instructions that help programmers get the most out of LRBni's extremely wide vector units. Starting with simple math examples that are fairly simple to vectorize, it moves through loops, conditionals, and more complex flow control, showing how to implement these algorithms in LRBni. Next, the numerous choices of data format are examined - when to use SOA or AOS (and what those terms mean!), and how to use gather/scatter most efficiently from the same data structures used in an existing engine. Finally, there is a quick look at efficient code scheduling and how to use the multiple hardware threads to help absorb instruction latencies. Presented by Tom Forsyth.
Procedural and Multi-core Techniques to Take Visuals to the Next Level
FRIDAY, MARCH 27
In this session we explore how Allegorithmic utilizes multi-core architectures in the Substance tools and middleware space to deliver unseen texture qualities and new production and rendering strategies to game teams. CPUs (especially mono-cores) are mainly utilized as control devices, massively undervaluing their adaptability and intrinsic power to perform more complex tasks. The rise of multi-cores will let middleware technologies (like Substance) utilize one or several of these cores to compute and modify textures on the fly. The immediate benefits of reduced download size, runtime footprint, reduced load times and scalability with platform are the just the beginning!
Substances opens up a world of new creative opportunities such as animation, interactive textures, asynchronous loading strategies, rich visual post-processing and rendering pipelines on the CPU side. Presented by Ed Plowman and Sebastian Deguy (Allegorithmic). |
|