Fast-Response and Low-Power Low-Dropout Regulator for In-Memory Computing

A low dropout (LDO) linear regulator with fast transient response and small overshoot voltage is proposed to address the problem of large and frequent quantization in existing large-scale in-memory computing (IMC) that seriously affects the stability of the reference voltage and the accuracy of the calculation. The regulator uses an enable circuit to save power in noncomputing states, a very low-voltage folding op-amp to achieve high gain, an overshoot voltage detection circuit to increase the swing rate current and thus improve the response speed, an adaptive bias circuit to provide bandwidth under heavy load and reduce power consumption under light load, and a PMOS power tube to achieve high drive current under low dropout voltage. The regulator is designed in a 65 nm standard CMOS process with a dropout voltage of 0.1 V and can provide a stable reference voltage with a maximum overshoot voltage of 30 mV and a maximum response time of 0.45 μs at 30 mA output current. The regulator is applied to a bit-serial large-scale IMC, and the results show that the IMC can read, write, and compute 256 bits of data quickly and stably, supporting 64 times more computing bits and 5.89 times faster recovery than existing buffer regulators. Compared with similar regulators, the drive capability is reduced, but the stable voltage can be provided at very low dropout voltage, which requires lower reference value and matches a wider range of IMC.

Authors

Institutions

Publication Details

Journal
Russian Microelectronics
Published
2024-12-01
DOI
https://doi.org/10.1134/s106373972460078x
Primary Topic
Advanced Memory and Neural Computing
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Fast-Response and Low-Power Low-Dropout Regulator for In-Memory Computing

Yang Liang, Mian Lou, Weidong Xu, Zongtai Li
Russian Microelectronics
Advanced Memory and Neural Computing
article

Fast-Response and Low-Power Low-Dropout Regulator for In-Memory Computing

Yang Liang, Mian Lou, Weidong Xu, Zongtai Li
article en

Abstract

A low dropout (LDO) linear regulator with fast transient response and small overshoot voltage is proposed to address the problem of large and frequent quantization in existing large-scale in-memory computing (IMC) that seriously affects the stability of the reference voltage and the accuracy of the calculation. The regulator uses an enable circuit to save power in noncomputing states, a very low-voltage folding op-amp to achieve high gain, an overshoot voltage detection circuit to increase the swing rate current and thus improve the response speed, an adaptive bias circuit to provide bandwidth under heavy load and reduce power consumption under light load, and a PMOS power tube to achieve high drive current under low dropout voltage. The regulator is designed in a 65 nm standard CMOS process with a dropout voltage of 0.1 V and can provide a stable reference voltage with a maximum overshoot voltage of 30 mV and a maximum response time of 0.45 μs at 30 mA output current. The regulator is applied to a bit-serial large-scale IMC, and the results show that the IMC can read, write, and compute 256 bits of data quickly and stably, supporting 64 times more computing bits and 5.89 times faster recovery than existing buffer regulators. Compared with similar regulators, the drive capability is reduced, but the stable voltage can be provided at very low dropout voltage, which requires lower reference value and matches a wider range of IMC.

Russian MicroelectronicsVol. 53(S1)
Beijing Microelectronics Technology Institute (CN)
Affordable and clean energy
Openalex Percentile: Top 74%
Advanced Memory and Neural Computing
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Fast-Response and Low-Power Low-Dropout Regulator for In-Memory Computing — Yang Liang, Mian Lou, et al. · Russian Microelectronics (2024) | TGRS Research Map | TGRS