Abstract
This paper studies a deferred pixel shading algorithm implemented on a Cell-based computer entertainment system. The pixel shader runs on the Synergistic Processing Units (SPUs) of the Cell and works concurrently with the GPU to render images. The system's unified memory architecture allows the Cell and GPU to exchange data through shared textures. The SPUs use the Cell DMA list capability to gather irregular fine-grained fragments of texture data generated by the GPU. They return resultant shadow textures the same way. The shading computation ran at up to 85 Hz at HDTV 720p resolution on 5 SPUs and generated 30.72 gigaops of performance. This is comparable to the performance of the algorithm running on a state of the art high end GPU. These results indicate that a hybrid solution in which the Cell and GPU work together can produce higher performance than either device working alone.

Speaker Bio
Alan Heirich is a senior software engineer with Sony Computer Entertainment America where he currently works on physics simulations for the PlayStation 3. He is the author of over 14 peer-reviewed publications, an editor of symposium proceedings for the IEEE, ACM and Eurographics, and currently serves as program co-chair for the ACM Sandbox Symposium on videogames. He has a PhD from the California Institute of Technology in Computer Science.

Abstract

Speaker Bio
Mike Acton is a Senior Architect working on PS3/Cell research at Highmoon Studios (Vivendi-Universal Games) who occasionally refers to himself in the third person. Most recently, Mike was the Lead Special Effects Programmer for Darkwatch at Highmoon Studios (previously Sammy Studios) in Carlsbad, CA. Previously Mike has held Lead Programming, Playstation 2 Engine Development and Research roles with Yuke's, VBlank and BlueSky/Titus. He worked for Sony Interactive Studios in PSX development. Mike has made regular appearances as a speaker at SCEA develpment conferences and has spoken at GDC. Mike Acton is not a framework-happy C++ programmer. He actually likes C. And assembly. In his spare time he develops hardware on FPGAs in VHDL. He prefers vi.
Mike also administers and writes for http://www.cellperformance.com, a techincal site devoted to performance issues on the Cell processor.

Abstract

Speaker Bio
Richard Lethin is President of Reservoir Labs, Inc. an independent systems research company with a focus area in advanced compilers, and PI on a research program to develop High Level Compilers within the DARPA Polymorphous Computer Architecture program.

Abstract
Cooperative use of SPUs is critical to the performance of multifaceted, real-time applications such as games. The SPU Threads model is necessary for sharing SPUs between processes, but has undesired overhead when sharing SPUs inside a process. Many systems have been designed for sharing SPUs among application tasks, but none of them perform optimally for all use cases. The SPU Runtime System (SPURS) solves this problem by introducing a common scheduling layer that can be shared by all existing and (hopefully) future SPU task systems. SPURS is a multi-layered, SPU-centric scheduling framework. At 2 kilobytes, the SPURS Kernel is the only binary that is always resident in Local Store. The SPURS Kernel is responsible for cooperative, coarse-grained scheduling of SPU Workloads. The contents of a Workload is interpreted and processed by its associated Policy Module. SPURS is provided with two Policy Modules to satisfy the most common use cases of SPUs: multi-tasking and job streaming. With support for custom Policy Modules, SPURS allows high performance interoperation of heterogeneous SPU workloads without imposing any specific programming model.

Speaker Bio

Abstract

Speaker Bio
Dr. Vivek Sarkar is Senior Manager of the Programming Technologies Department in the IBM Research Division, and is a member of the IBM Academy of Technology.

1. The projects under way in his department span the areas of:
2. Programming Models --- XJ, PERCS/X10, Symbiotic
3. Programming Tools --- Advanced Refactoring, C/C++ Grid Migration, Concern Manipulation Environment (CME), J2EE Validation and Optimization (DOMO), Legacy Transformation (Mastery), Test Prioritization (Chianti)
4. Software Virtualization and Delivery (PDS)
5. Advanced Virtual Machine Technologies (Jikes RVM, Parley, Memory Wall)

Dr. Sarkar is also responsible for coordinating Eclipse-related projects and activities in IBM Research, including the world-wide Eclipse Innovation Grants program. Dr. Sarkar's personal research interest is in advancing the state of the art of programming language design, programming tools, and optimizing and parallelizing compilers so as to deliver improved productivity and performance benefits to users of parallel systems. He is the leader of the PERCS Programming Model and Tools project, which includes work on X10, an experimental object-oriented programming language and programming environment being developed to address the locality and scalability challenges facing parallel systems in the 2010 timeframe. He has worked on several compiler-related projects at IBM including the PTRAN and Jikes RVM research projects at Watson, and high-order transformations and OpenMP parallelization in the IBM XL Fortran product compilers. Dr. Sarkar has published over 75 papers in the area of optimizing and parallelizing compilers, and has lectured internationally on his research. Dr. Sarkar joined IBM in 1987, after obtaining a Ph.D. from Stanford University. In 1997, he was on sabbatical as a visiting associate professor at MIT.