Bulldozers, Termites, and Chainsaws: Finding a Gradual Path to Parallelism


October 4, 2007


The transition to multicore is effectively a gamble that application programmers will start writing all of their code in parallel, or that workloads will change to make it simple to generate many threads. A key question for architects is the granularity of the processors on each multicore chip – many small processors, a few medium-sized processors, or fewer large behemoths. In this talk, I will show that with an EDGE instruction set, this choice is a false choice. I will describe Composable Lightweight Processors, which allow larger logical processors to be synthesized out of smaller ones. This class of architectures treats large processors as a distributed system made up of many small pieces on an on-chip micronetwork. This approach creates the opportunity to have many intermediate design points between traditional uniprocessors and multiprocessors, blurring the distinction and permitting a more gradual path to fine-grain parallel execution than requiring full-blown parallel applications immediately.


Doug Burger

Doug Burger is an Associate Professor of Computer Sciences at the University of Texas at Austin. He received his B.S. in Computer Science from Yale University in 1991, and his M.S. and Ph.D. in Computer Sciences from the University of Wisconsin-Madison in 1993 and 1998, respectively. His research interests include computer architecture, supercomputing, compiler design and implementation, as well as the implications of novel computing technologies. He co-leads the TRIPS project at UT-Austin, which has developed a full system prototype of a new class of instruction-set architectures called Explicit Data Graph Execution (or EDGE). He is a Senior Member of the IEEE and ACM, Chair of ACM SIGARCH, and was the recipient of the 2006 ACM Maurice Wilkes Award.