A Multicore Neuromorphic Chip Design Using Multilevel Synapses in 65nm Standard CMOS Technology
2015-06
Loading...
View/Download File
Persistent link to this item
Statistics
View StatisticsJournal Title
Journal ISSN
Volume Title
Title
A Multicore Neuromorphic Chip Design Using Multilevel Synapses in 65nm Standard CMOS Technology
Authors
Published Date
2015-06
Publisher
Type
Thesis or Dissertation
Abstract
Neuromorphic research community is focused on designing a hardware which is as efficient as biological brain in terms of performance, power and area. It opens up opportunities to optimize these designs at all levels from architecture to devices. We propose a novel architecture to have tight integration between neurons and synapses. Our 32K bit neuromorphic chip with 256 axons and 256 neurons demonstrates 4 neuromorphic cores operating in a completely parallel fashion. Eflash memory core representing synapses saves power and area. The Non-volatility of eflash consumes zero static power. The ability to store multi-levels of weights in a single cell makes the array denser. Unlike flash technology, eflash doesn’t require specialized fabrication process, hence the neuromorphic chip is implemented in 65nm standard CMOS technology. The current sensing neurons with parallel reading scheme makes the neuronal operation several orders of magnitude faster than state-of-the-art neuromorphic designs.
Keywords
Description
University of Minnesota M.S. thesis. June 2015. Major: Electrical Engineering. Advisor: Chris Kim. 1 computer file (PDF); v, 38 pages.
Related to
Replaces
License
Series/Report Number
Funding information
Isbn identifier
Doi identifier
Previously Published Citation
Other identifiers
Suggested citation
Tagare, Deepak Kumar. (2015). A Multicore Neuromorphic Chip Design Using Multilevel Synapses in 65nm Standard CMOS Technology. Retrieved from the University Digital Conservancy, https://hdl.handle.net/11299/190620.
Content distributed via the University Digital Conservancy may be subject to additional license and use restrictions applied by the depositor. By using these files, users agree to the Terms of Use. Materials in the UDC may contain content that is disturbing and/or harmful. For more information, please see our statement on harmful content in digital repositories.