cs 8501 networks-on-chip ( nocs )
DESCRIPTION
CS 8501 Networks-on-Chip ( NoCs ). Lukasz Szafaryn 15 FEB 10. Motivation. Bus has been the most popular interconnect for multiprocessor systems When scaling feature sizes and frequency, wire delays remain larger than clock cycle. - PowerPoint PPT PresentationTRANSCRIPT
![Page 1: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/1.jpg)
CS 8501Networks-on-Chip (NoCs)
Lukasz Szafaryn15 FEB 10
![Page 2: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/2.jpg)
Motivation• Bus has been the most popular interconnect for
multiprocessor systems• When scaling feature sizes and frequency, wire delays
remain larger than clock cycle
• Need for interconnect with deterministic delays and scalability
• When expanding to a many core-system, contention decreases throughput
![Page 3: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/3.jpg)
What is a Network-on-Chip (NoC)?• Leveraging existing computer networking
principles to improve inter-component intra-chip communications
• Each on-chip component connected by an intelligent switch to particular communication wire(s)
• Improvement over standard bus based interconnections for SoC architectures in terms of throughput
![Page 4: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/4.jpg)
Topologies
CLICHÉ Torus Folded torus
![Page 5: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/5.jpg)
Topologies contd.
SPIN Octagon BFT
![Page 6: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/6.jpg)
SwitchingCircuit Switching• Dedicated path, or circuit, is established over which data packets will
travel • Naturally lends itself to time-sensitive guaranteed service due to
resource allocation• Reservation of bandwidth decreases overall throughput and increases
average delays
Packet Switching• Intermediate routers are now responsible for the routing of individual
packets through the network, rather than following a single path• Provides for so-called best-effort services• Sharing of resources allows for higher throughput
![Page 7: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/7.jpg)
Switching contd.Wormhole Switching• Message is divided up into smaller, fixed length flow units called flits• Only first flit contains routing information, subsequent flits follow• Buffer size is significantly reduced due to the limitation on the number
of flits needed to be buffered at any given time
Virtual Channels• Allows for several instances of wormhole switching• Additional buffers are added, which increases overall switch size, but
significantly increases throughput
![Page 8: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/8.jpg)
Performance Metrics
Simulator developed to measure:• Throughput (in flits)• Latency (of flits)• Energy (per packet)
Hardware model developed to estimate:• Area (router and link overhead)
![Page 9: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/9.jpg)
Simulator Setup
• SoC with 256 fixed-size (100K gate) elements
• Topologies use the same type of switching (wormhole with usually 4 virtual channels)
• Topologies use different routing schemes (different configurations of switches and links)
![Page 10: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/10.jpg)
Simulated Traffic• Uniform and localized traffic patterns
• Poisson and self-similar injection methods
![Page 11: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/11.jpg)
Number of Virtual Channels
Throughput Latency
![Page 12: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/12.jpg)
Number of Virtual Channels contd.
Energy Dissipation
![Page 13: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/13.jpg)
Injection Load
AcceptedTraffic
Latency
![Page 14: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/14.jpg)
Injection Load contd.
EnergyDissipation
![Page 15: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/15.jpg)
Localization
Throughput
![Page 16: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/16.jpg)
Localization contd.
Latency(30% local traffic)
Latency(80% local traffic)
![Page 17: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/17.jpg)
Localization contd.
EnergyDissipation
(30% local traffic)
EnergyDissipation
(80% local traffic)
![Page 18: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/18.jpg)
Interconnect Area
Die size: 20mm x 20mm
![Page 19: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/19.jpg)
Interconnect Area contd.
![Page 20: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/20.jpg)
Case Study
![Page 21: CS 8501 Networks-on-Chip ( NoCs )](https://reader033.vdocuments.site/reader033/viewer/2022051019/5681598a550346895dc6cd52/html5/thumbnails/21.jpg)
Conclusions
• NoC is an interconnect architecture of choice of multiprocessor SoC and many-core systems
• The main trade-off is between performance and energy
• Other considerations could lead to different energy/performance numbers