Scaling with Design Constraints: Predicting the Future of Big Chips
Authors:
- 1 July 2011
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Micro
- Vol. 31 (4) , 16-29
Abstract
The past few years have witnessed high-end processors with increasing numbers of cores and larger dies. With limited instruction-level parallelism, chip power constraints, and technology-scaling limitations, designers have embraced multiple cores rather than single-core performance scaling to improve chip throughput. This article examines whether this approach is sustainable by scaling from a state-of-the-art big-chip design point using analytical models.Keywords
This publication has 17 references indexed in Scilit:
- The Exascale challenge2010
- The implementation of POWER7TM: A highly parallel and scalable multi-core high-end server processor2010
- Architecting for power management: The IBM® POWER7™ approach2010
- ORION 2.0: A fast and accurate NoC power and area model for early-stage design space exploration2009
- Embedded Memories for Nano-Scale VLSIs2009
- Many-core design from a thermal perspective2008
- An 80-Tile Sub-100-W TeraFLOPS Processor in 65-nm CMOSIEEE Journal of Solid-State Circuits, 2008
- Power-constrained high-frequency circuits for the IBM POWER6 microprocessorIBM Journal of Research and Development, 2007
- Thousand Core ChipsA Technology PerspectiveProceedings of the 39th conference on Design automation - DAC '02, 2007
- Temperature-aware microarchitectureACM Transactions on Architecture and Code Optimization, 2004