SC Conference - Activity Details

Node Level Parallelism for Large Scale Supercomputers


James C. Browne  (University of Texas at Austin)
Keshav Pingali  (University of Texas at Austin)
Workshops Session
Sunday,  08:30AM - 05:00PM
Room 14
This workshop focuses specifically on parallelism for multichip, multicore nodes of supercomputers. The goal is to gather the current state of knowledge spanning programming and execution models and scheduling algorithms and to articulate effective approaches spanning multiple configurations and application communities. The workshop will have a core of invited participants but will be open for public participation with limited enrollment. Workshop leadership will be by a committee recognized research leaders. Software development for multicore processors is widely regarded as one of the most challenging problems facing the computer systems community. Today's supercomputers contain tens of thousands of multicore processors, so the task of programming these machines is even more difficult: in particular, node-level parallelism to exploit multicore features must be integrated into scalable global parallelism. The experience gained in parallelism for this generation of supercomputer nodes is highly relevant to future generations of multicore chip desktop and server systems.
   IEEE Computer Society  /  ACM     2 0   Y E A R S   -   U N L E A S H I N G   T H E   P O W E R   O F   H P C