System Performance: From Enterprise to AI
Authors:
- 1 December 2018
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 28.1.1-28.1.4
Abstract
System performance has shown many decades of continuous improvement. After first reviewing historical trends and the current outlook, in this paper we discuss the challenges and opportunities in future computing systems due to the disruptive confluence of stalled scaling and emerging AI workloads. Heterogeneous integration is highlighted as a key means to future systems performance growth.Keywords
This publication has 9 references indexed in Scilit:
- Equivalent-accuracy accelerated neural-network training using analogue memoryNature, 2018
- Recent Advances and New Trends in Flip Chip TechnologyJournal of Electronic Packaging, 2016
- PECVD low and ultralow dielectric constant materials: From invention and research to productsJournal of Vacuum Science & Technology B, 2016
- Heterogeneous Integration for Performance and ScalingIEEE Transactions on Components, Packaging and Manufacturing Technology, 2016
- Practical Strategies for Power-Efficient Computing TechnologiesProceedings of the IEEE, 2010
- RooflineCommunications of the ACM, 2009
- Design of micron MOS switching devicesIEEE Solid-State Circuits Society Newsletter, 2007
- Silicon CMOS devices beyond scalingIBM Journal of Research and Development, 2006
- High-Performance Throughput ComputingIEEE Micro, 2005