BEGIN:VCALENDAR PRODID:-//Microsoft Corporation//Outlook MIMEDIR//EN VERSION:1.0 BEGIN:VEVENT DTSTART:20111113T163000Z DTEND:20111114T010000Z LOCATION: DESCRIPTION;ENCODING=QUOTED-PRINTABLE:ABSTRACT: Developers need a comprehensive set of strategies and technologies for tackling parallel programming. Intel Parallel Building Blocks is such a comprehensive set of technologies, and includes Threading Building Blocks (now extremely popular), Cilk (from MIT), and Array Building Blocks (from RapidMind). PBB supports a set of sophisticated and structured parallel programming strategies. Many patterns of parallel computation are supported from fork-join parallelism to collective vector operations. PBB supports multiple hardware mechanisms for performance through parallelism, including vectorization (SIMD instructions), multicore, and co-processor offload. Static and dynamic compilation are available.=0A=0AWe introduce PBB and demonstrate detailed usage through a number of examples. Examples cover a range of algorithmic patterns and application domains, including scientific computation, finance, and imaging. Since PBB integrates a number of different programming paradigms, multiple approaches to solving the same problems are compared. Includes live demonstrations of build processes, performance and scalability analysis, and performance optimization. SUMMARY:S15: Understanding Intel Parallel Building Blocks PRIORITY:3 END:VEVENT END:VCALENDAR