DocumentCode
3694087
Title
A hexagonal shaped processor and interconnect topology for tightly-tiled many-core architecture
Author
Zhibin Xiao;Bevan Baas
Author_Institution
Department of Electrical and Computer Engineering, University of California, Davis, USA
fYear
2012
Firstpage
153
Lastpage
158
Abstract
2-Dimensional meshes are the most commonly used Network-on-Chip (NoC) topology for on-chip communication in many-core processor arrays due to their low complexity and excellent match to rectangular processor tiles. However, 2D meshes may incur local traffic congestion for applications with significant levels of traffic with non-neighboring cores, resulting in long latencies and high power consumption. In this paper, we propose an 8-neighbor mesh topology and a 6-neighbor topology with hexagonal-shaped processor tiles. A 16-bit DSP processor and the corresponding processor arrays are implemented in all three topologies. The hexagonal processor tile and arrays of tiles are laid out using industry-standard CAD tools and automatic place and route flow without full-custom design, and result in DRC-clean and LVS-clean layout. A 1080p H.264/AVC residual video encoder and a 54 Mbps 802.11a/11g OFDM wireless LAN baseband receiver are mapped onto all topologies. The 6-neighbor hexagonal grid topology incurs a 2.9% area increase per tile compared to the 4-neighbor 2D mesh, but its much more effective inter-processor interconnect yields an average total application area reduction of 21%, an average power reduction of 17%, and a total application inter-processor communication distance reduction of 19%.
Keywords
"Topology","Routing","Legged locomotion"
Publisher
ieee
Conference_Titel
VLSI and System-on-Chip, 2012 (VLSI-SoC), IEEE/IFIP 20th International Conference on
Print_ISBN
978-1-4673-2658-2
Type
conf
DOI
10.1109/VLSI-SoC.2012.7332093
Filename
7332093
Link To Document