Magnetic domain wall based synaptic and activation function generator for neuromorphic accelerators
Magnetic domain walls are information tokens in both logic and memory devices and hold particular interest in applications such as neuromorphic accelerators that combine logic in memory. Here, we show that devices based on the electrical manipulation of magnetic domain walls are capable of implement...
| Main Authors: | , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
American Chemical Society (ACS)
2020
|
| Online Access: | https://hdl.handle.net/1721.1/127775 |
| _version_ | 1826211036355624960 |
|---|---|
| author | Siddiqui, Saima Afroz Dutta, Sumit Tang, Astera S. Liu, Luqiao Ross, Caroline A. Baldo, Marc A |
| author2 | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science |
| author_facet | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Siddiqui, Saima Afroz Dutta, Sumit Tang, Astera S. Liu, Luqiao Ross, Caroline A. Baldo, Marc A |
| author_sort | Siddiqui, Saima Afroz |
| collection | MIT |
| description | Magnetic domain walls are information tokens in both logic and memory devices and hold particular interest in applications such as neuromorphic accelerators that combine logic in memory. Here, we show that devices based on the electrical manipulation of magnetic domain walls are capable of implementing linear, as well as programmable nonlinear, functions. Unlike other approaches, domain-wall-based devices are ideal for application to both synaptic weight generators and thresholding in deep neural networks. Prototype micrometer-size devices operate with 8 ns current pulses and the energy consumption required for weight modulation is ≤16 pJ. Both speed and energy consumption compare favorably to other synaptic nonvolatile devices, with the expected energy dissipation for scaled 20 nm devices close to that of biological neurons. |
| first_indexed | 2024-09-23T14:59:33Z |
| format | Article |
| id | mit-1721.1/127775 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T14:59:33Z |
| publishDate | 2020 |
| publisher | American Chemical Society (ACS) |
| record_format | dspace |
| spelling | mit-1721.1/1277752022-09-29T11:55:10Z Magnetic domain wall based synaptic and activation function generator for neuromorphic accelerators Siddiqui, Saima Afroz Dutta, Sumit Tang, Astera S. Liu, Luqiao Ross, Caroline A. Baldo, Marc A Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology. Department of Materials Science and Engineering Magnetic domain walls are information tokens in both logic and memory devices and hold particular interest in applications such as neuromorphic accelerators that combine logic in memory. Here, we show that devices based on the electrical manipulation of magnetic domain walls are capable of implementing linear, as well as programmable nonlinear, functions. Unlike other approaches, domain-wall-based devices are ideal for application to both synaptic weight generators and thresholding in deep neural networks. Prototype micrometer-size devices operate with 8 ns current pulses and the energy consumption required for weight modulation is ≤16 pJ. Both speed and energy consumption compare favorably to other synaptic nonvolatile devices, with the expected energy dissipation for scaled 20 nm devices close to that of biological neurons. National Science Foundation (U.S.) (Award 1639921) 2020-09-30T14:30:22Z 2020-09-30T14:30:22Z 2019-12 2020-09-10T17:50:20Z Article http://purl.org/eprint/type/JournalArticle 1530-6984 https://hdl.handle.net/1721.1/127775 Siddiqui, Saima A. et al. “Magnetic domain wall based synaptic and activation function generator for neuromorphic accelerators.” Nano Letters, 20, 2 (December 2019): 1033–1040 © 2019 The Author(s) en 10.1021/ACS.NANOLETT.9B04200 Nano Letters Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use. application/pdf American Chemical Society (ACS) arXiv |
| spellingShingle | Siddiqui, Saima Afroz Dutta, Sumit Tang, Astera S. Liu, Luqiao Ross, Caroline A. Baldo, Marc A Magnetic domain wall based synaptic and activation function generator for neuromorphic accelerators |
| title | Magnetic domain wall based synaptic and activation function generator for neuromorphic accelerators |
| title_full | Magnetic domain wall based synaptic and activation function generator for neuromorphic accelerators |
| title_fullStr | Magnetic domain wall based synaptic and activation function generator for neuromorphic accelerators |
| title_full_unstemmed | Magnetic domain wall based synaptic and activation function generator for neuromorphic accelerators |
| title_short | Magnetic domain wall based synaptic and activation function generator for neuromorphic accelerators |
| title_sort | magnetic domain wall based synaptic and activation function generator for neuromorphic accelerators |
| url | https://hdl.handle.net/1721.1/127775 |
| work_keys_str_mv | AT siddiquisaimaafroz magneticdomainwallbasedsynapticandactivationfunctiongeneratorforneuromorphicaccelerators AT duttasumit magneticdomainwallbasedsynapticandactivationfunctiongeneratorforneuromorphicaccelerators AT tangasteras magneticdomainwallbasedsynapticandactivationfunctiongeneratorforneuromorphicaccelerators AT liuluqiao magneticdomainwallbasedsynapticandactivationfunctiongeneratorforneuromorphicaccelerators AT rosscarolinea magneticdomainwallbasedsynapticandactivationfunctiongeneratorforneuromorphicaccelerators AT baldomarca magneticdomainwallbasedsynapticandactivationfunctiongeneratorforneuromorphicaccelerators |