IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL. 56, NO. 4, APRIL 2021 1019
Introduction to the Special Issue on the
2020 Symposium on VLSI Circuits
T
HIS Special Issue of the IEEE JOURNAL OF SOLID-
S
TATE CIRCUITS highlights some of the best papers
presented at the Symposium on VLSI Circuits, held on
June 15–19, 2020. Due to the Covid-19 pandemic, this was
the first VLSI Symposium that was held fully virtual. While it
is difficult to replace all of the interactions of an in-person con-
ference, the symposium featured all of the technical material,
including regular and invited papers, plenaries, short courses,
and a forum. In addition, new content was added, including
a series of executive sessions where paper authors, technical
program committee members, and the audience engaged in
open discussions on key topics related to our field. While the
live events were concentrated in that one week, on demand
content was available through the end of August for partici-
pants to view at their leisure.
Founded in 1987, the Symposium on VLSI Circuits is
jointly sponsored by the IEEE Solid-State Circuits Society
and the Japan Society of Applied Physics in cooperation
with the Institute of Electronics, Information, and Commu-
nications Engineers and the IEEE Electron Devices Society.
The Symposium on VLSI Circuits is jointly organized and
colocated with the Symposium on VLSI Technology and
thus provides excellent opportunities for interactions among
integrated circuit and technology experts on topics of mutual
interest. As the new normal emerges, the VLSI Symposia are
committed to developing new ways to deliver content and to
providing enhanced opportunities for interaction.
The 2020 Symposium on VLSI Circuits received 321 sub-
missions. The program committees selected 110 papers for
presentation with subjects including digital circuits, proces-
sors, SoCs, machine learning accelerators, memories, biomed-
ical circuits, sensors and displays, power conversion circuits,
analog amplifiers and filters, wireless and wireline commu-
nications, data converters, frequency generation, and clock
circuits. This issue of the journal includes 26 outstanding
articles selected as highlights of the work presented at the
2020 Symposium. The journal articles describe the work
introduced at the symposium with greater detail than pro-
vided by corresponding papers in the symposium digest. The
articles were subject to the standard journal review process.
We enjoyed working with the authors on these articles, and
we certainly hope that the technical details presented in these
articles will be interesting and useful to journal readers. The
following paragraphs give an overview regarding the article
content of this Special Issue.
The first article presents a robust light detection and rang-
ing (LiDAR) system with an embedded interference filter.
Digital Object Identifier 10.1109/JSSC.2021.3059008
Researchers from Sungkyunkwan University, SOS LAB, and
Samsung Electronics demonstrate a reliable 36-channel direct
time-of-flight measurement even when 32 different LiDARs
interfere.
The next three articles advance sensor front-ends. For
safety-critical touch applications, the Delft University of Tech-
nology and Cypress Semiconductor achieve a low-power eddy-
current sensor interface with a power consumption of 200 μW
from 1.8-V power supply. Authors from the University of
California at San Diego (UCSD) achieve a 112-dB spurious
free dynamic range (SFDR) and 89-dB signal-to-noise and dis-
tortion ratio (SNDR) voltage control oscillator (VCO)-based
sensor front-end. The University of Virginia presents a 785-nW
multimodal sensor interface IC for ozone pollutant sensing
and correlated cardiovascular disease monitoring based on
electrocardiography (ECG) and photoplethysmography (PPG).
With the advancement of deep neural networks (DNNs),
a more efficient processing architecture is desired. A DNN-
based image signal processor presented by the Univer-
sity of Michigan achieves hierarchical image recognition
with 170 μW at 5 f/s. Authors from Intel demonstrate a
617 TOPS/W digital binary neural network accelerator in
10-nm FinFET CMOS. An energy-efficient processing-in-
memory (PIM) architecture proposed by the Korea Advanced
Institute of Science and Technology achieves more than
2.1 times improvement over the state-of-the-art PIM imple-
mentations. Researchers from the University of Michigan
and Taiwan Semiconductor Manufacturing Company (TSMC)
also demonstrate an energy-efficient DNN accelerator fea-
turing 24-Mbit random access memory (RAM) as all-on-
chip weight storage to eliminate energy-consuming off-chip
memory accesses.
The next two articles showcase the state-of-the-art memory
implementations. A dual-port spin-orbit-torque magnetore-
sistive random access memory (SOT-MRAM) presented by
Tohoku University and JST ACCEL achieves 90-MHz read
and 60-MHz write operations under field-assistance-free con-
ditions. Samsung Electronics shows a 1.2 V, 1.8 Gb/s/pin, and
16-Tb
NAND flash memory multi-chip package incorporating
16 dies of 1-Tb
NAND flash memory each.
To address the demands of increasing security, Intel demon-
strates a side-channel attack (SCA) hardened AES-128 and
RSA cryptoprocessor in 14-nm CMOS using a nonlinear dig-
ital low-dropout regulator with control loop randomizations.
Two microprocessor articles advance the processor
architecture to improve energy efficiency. Yokohama National
University demonstrates the first successful adiabatic
microprocessor using 1.4-zJ/op Josephson junction devices
manufactured using an Nb/AlOx/Nb superconductor process.
Researchers from Qualcomm Technologies and The University
0018-9200 © 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.