Unveiling the UT-Austin TRIPS Processor: An Architecture for Scaling to the End of Silicon


Monday, April 30, 2007, 5 - 6 p.m. ACES Avaya Auditorium, Room 2.302

TRIPS (The Tera-op, Reliable, Intelligently adaptive Processing System) is a revolutionary new polymorphous microprocessor architecture designed and built by the Department of Computer Sciences at The University of Texas at Austin. For the past seven years the research team, led by Professors Doug Burger, Stephen Keckler and Kathryn McKinley, has been working on the design of the processor, an updated cross-platform compiler and the instruction set architecture needed to run it.

The team's goal is to produce a scalable architecture that can accelerate industrial, consumer, embedded and scientific workloads, reaching trillions of calculations per second on a single chip. This radical new architecture can produce improved single-thread performance-at greater power efficiencies-than conventional designs. This technology, called Explicit Data Graph Execution (EDGE architectures) offers a flexible alternative to the current industrial direction of providing a greater number of processor cores with each passing generation.

The TRIPS prototype processor is a working system that demonstrates the capabilities of this new EDGE technology. Its architecture is composed of many copies of a small number of replicated tiles, reducing complexity and improving ease of design. Produced with generous support from DARPA, each TRIPS prototype processor contains two processing cores, each of which can issue 16 operations per cycle with up to 1,024 instructions in flight simultaneously. The prototype is the first on a roadmap that will lead to ultra-powerful, flexible processors implemented in nanoscale technologies. This technology is expected to have significant impact on the computing industry.

Comments

Post new comment

The content of this field is kept private and will not be shown publicly.

Comments are moderated. They will be posted if they stick to the topic and contribute to the conversation. They will not be published if they contain or link to abusive material, personal attacks, profanity or spam.