Achieving High Core Neuron Density in a Neuromorphic Chip Through Trade-off Among Area, Power Consumption, and Data Access Bandwidth
As a crucial component of neuromorphic chips, on-chip memory usually occupies most of the on-chip resources and limits the improvement of neuron density. The alternative of using off-chip memory may result in additional power consumption or even a bottleneck for off-chip data access. This article proposes an on- and off-chip co-design approach and a figure of merit (FOM) to achieve a trade-off between chip area, power consumption, and data access bandwidth. By evaluating the FOM of each design scheme, the scheme with the highest FOM (1.085× better than the baseline) is adopted to design a neuromorphic chip. Deep multiplexing and weight-sharing technologies are used to reduce on-chip resource overhead and data access pressure. A hybrid memory design method is proposed to optimize on- and off-chip memory distribution, which reduces on-chip storage pressure and total power consumption by 92.88% and 27.86%, respectively, while avoiding the explosion of off-chip access bandwidth. The co-designed neuromorphic chip with ten cores fabricated under standard 55 nm CMOS technology has an area of 4.4 mm 2 and a core neuron density of 4.92 K/mm 2, an improvement of 3.39 ∼ 30.56× compared with previous works. After deploying a full-connected and a convolution-based spiking neural network (SNN) for ECG signal recognition, the neuromorphic chip achieves 92% and 95% accuracy, respectively. This work provides a new path for developing high-density and large-scale neuromorphic chips.
Medienart: |
E-Artikel |
---|
Erscheinungsjahr: |
2023 |
---|---|
Erschienen: |
2023 |
Enthalten in: |
Zur Gesamtaufnahme - volume:17 |
---|---|
Enthalten in: |
IEEE transactions on biomedical circuits and systems - 17(2023), 6 vom: 21. Dez., Seite 1319-1330 |
Sprache: |
Englisch |
---|
Beteiligte Personen: |
Zhou, P J [VerfasserIn] |
---|
Links: |
---|
Themen: |
---|
Anmerkungen: |
Date Completed 11.01.2024 Date Revised 11.01.2024 published: Print-Electronic Citation Status MEDLINE |
---|
doi: |
10.1109/TBCAS.2023.3292469 |
---|
funding: |
|
---|---|
Förderinstitution / Projekttitel: |
|
PPN (Katalog-ID): |
NLM359070531 |
---|
LEADER | 01000caa a22002652 4500 | ||
---|---|---|---|
001 | NLM359070531 | ||
003 | DE-627 | ||
005 | 20240114233730.0 | ||
007 | cr uuu---uuuuu | ||
008 | 231226s2023 xx |||||o 00| ||eng c | ||
024 | 7 | |a 10.1109/TBCAS.2023.3292469 |2 doi | |
028 | 5 | 2 | |a pubmed24n1256.xml |
035 | |a (DE-627)NLM359070531 | ||
035 | |a (NLM)37405896 | ||
040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
041 | |a eng | ||
100 | 1 | |a Zhou, P J |e verfasserin |4 aut | |
245 | 1 | 0 | |a Achieving High Core Neuron Density in a Neuromorphic Chip Through Trade-off Among Area, Power Consumption, and Data Access Bandwidth |
264 | 1 | |c 2023 | |
336 | |a Text |b txt |2 rdacontent | ||
337 | |a ƒaComputermedien |b c |2 rdamedia | ||
338 | |a ƒa Online-Ressource |b cr |2 rdacarrier | ||
500 | |a Date Completed 11.01.2024 | ||
500 | |a Date Revised 11.01.2024 | ||
500 | |a published: Print-Electronic | ||
500 | |a Citation Status MEDLINE | ||
520 | |a As a crucial component of neuromorphic chips, on-chip memory usually occupies most of the on-chip resources and limits the improvement of neuron density. The alternative of using off-chip memory may result in additional power consumption or even a bottleneck for off-chip data access. This article proposes an on- and off-chip co-design approach and a figure of merit (FOM) to achieve a trade-off between chip area, power consumption, and data access bandwidth. By evaluating the FOM of each design scheme, the scheme with the highest FOM (1.085× better than the baseline) is adopted to design a neuromorphic chip. Deep multiplexing and weight-sharing technologies are used to reduce on-chip resource overhead and data access pressure. A hybrid memory design method is proposed to optimize on- and off-chip memory distribution, which reduces on-chip storage pressure and total power consumption by 92.88% and 27.86%, respectively, while avoiding the explosion of off-chip access bandwidth. The co-designed neuromorphic chip with ten cores fabricated under standard 55 nm CMOS technology has an area of 4.4 mm 2 and a core neuron density of 4.92 K/mm 2, an improvement of 3.39 ∼ 30.56× compared with previous works. After deploying a full-connected and a convolution-based spiking neural network (SNN) for ECG signal recognition, the neuromorphic chip achieves 92% and 95% accuracy, respectively. This work provides a new path for developing high-density and large-scale neuromorphic chips | ||
650 | 4 | |a Journal Article | |
700 | 1 | |a Zuo, Y |e verfasserin |4 aut | |
700 | 1 | |a Qiao, G C |e verfasserin |4 aut | |
700 | 1 | |a Zhang, C M |e verfasserin |4 aut | |
700 | 1 | |a Zhang, Z |e verfasserin |4 aut | |
700 | 1 | |a Meng, L W |e verfasserin |4 aut | |
700 | 1 | |a Yu, Q |e verfasserin |4 aut | |
700 | 1 | |a Liu, Y |e verfasserin |4 aut | |
700 | 1 | |a Hu, S G |e verfasserin |4 aut | |
773 | 0 | 8 | |i Enthalten in |t IEEE transactions on biomedical circuits and systems |d 2007 |g 17(2023), 6 vom: 21. Dez., Seite 1319-1330 |w (DE-627)NLM192068652 |x 1940-9990 |7 nnns |
773 | 1 | 8 | |g volume:17 |g year:2023 |g number:6 |g day:21 |g month:12 |g pages:1319-1330 |
856 | 4 | 0 | |u http://dx.doi.org/10.1109/TBCAS.2023.3292469 |3 Volltext |
912 | |a GBV_USEFLAG_A | ||
912 | |a GBV_NLM | ||
951 | |a AR | ||
952 | |d 17 |j 2023 |e 6 |b 21 |c 12 |h 1319-1330 |