SHARCNET-IBM Workshop on Cell Computing

The Cell BE (Cell) architecture incorporates several interesting features which enable Cell to support highly parallel, compute intensive codes. The first implementation of the architecture results a single chip of heterogeneous processors, including a POWER processing element PPE with two levels of cache and eight synergistic processing elements SPEs with their own local memories and globally consistent DMA engines. In addition to processor-level parallelism, each processing element has Single Instruction Multiple Data (SIMD) units that can each process from 4 words up to 16 chars per processing cycle. Such design allows Cell to support a wide variety of high performance workloads across a number of industries including digital media, medical imaging, aerospace, defense and communications, and financial services. While PlayStation 3 is the first major piece of hardware to use Cell in the gaming industry, the Roadrunner project deployed at Los Alamos National Laboratory in May 2008 is the first large-scale heterogeneous supercomputer system deployment containing 16,000 Cell processors hybridizing with 16,000 Opteron X64 delivering 1.1 petaflops. Programming techniques which harness the power of Cell are key to attaining the high compute performance of which it is capable, but such techniques may require some different programming concepts. The first day of this two-day Cell BE programming workshop is designed for beginners and provides a technical overview of the Cell architecture, programming models, and software development environment. It will also cover the IBM SDK (Software Development Toolkit) for Multicore Acceleration Version 3.0 plus hands-on exercise to give participants a jump start opportunity to practice Cell programming.

The second day of the course is primarily designed for the more advanced students who have some programming experience with the SDK and the Cell BE Simulator (these students are strongly recommended to join the first day class to re-familiar themselves with the fundamentals). It will address problems such as what are the performance bottlenecks in Cell program? how should the bottlenecks be handled? how should Cell programmers SIMDize their code? how should they parallelize it? how the codes be synchronized? and processes be communicated? etc.

The two day course will be taught by two instructors from IBM, Dr. Michael Perrone and Dr. Robert Enenkel.

The workshop is open to all, registration is required.