CoChat

A 64-Tile 2.4-Mb In-Memory-Computing CNN Accelerator Employing Charge-Domain Compute

Valavi, Hossein, Ramadge, Peter J., Nestler, Eric, Verma, Naveen · 2019 · IEEE Journal of Solid-State Circuits

The DOI matches its Crossref record; 54 of the 61 references it cites were found in a registry or an index.

Abstract

Large-scale matrix-vector multiplications, which dominate in deep neural networks (DNNs), are limited by data movement in modern VLSI technologies. This paper addresses data movement via an in-memory-computing accelerator that employs charged-domain mixed-signal operation for enhancing compute SNR and, thus, scalability. The architecture supports analog/binary input activation (IA)/weight first layer (FL) and binary/binary IA/weight hidden layers (HLs), with batch normalization and input-output (IO) (buffering) circuitry to enable cascading, if desired, for realizing different DNN layers. The architecture is arranged as 8 × 8 = 64 in-memory-computing neuron tiles, supporting up to 512, 3 × 3 × 512-input HL neurons and 64, 3 × 3 × 3-input FL neurons, configurable via tile-level clock gating. In-memory computing is achieved using an 8T bit cell with overlaying metal-oxide-metal (MOM) capacitor, yielding a structure having 1.8 × the area of a standard 6T bit cell. Implemented in 65-nm CMOS, the design achieves HLs/FL energy efficiency of 866/1.25 TOPS/W and throughput of 18876/43.2 GOPS (1498/3.43 GOPS/mm2), when implementing convolution layers; and 658/0.95 TOPS/W, 9438/10.47 GOPS (749/0.83 GOPS/mm2), when implementing convolution followed by batch normalization layers. Several large-scale neural networks are demonstrated, showing performance on standard benchmarks (MNIST, CIFAR-10, and SVHN) equivalent to ideal digital computing.

Overview

Integrity signals

Registry

Crossref · 10.1109/jssc.2019.2899730

Cites

54 of 54 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Verification

Claims the paper makes

This paper’s full text isn’t openly available, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Show all 54 references

Research works printed but not linked: 4

  1. jia, 2018, A microprocessor implemented in 65nm CMOS with configurable and bit-scalable accelerator for programmable in-memory computing
  2. bang, 2017, A $288\mu\text{W}$ programmable deep-learning processor with 270kB on-chip weight storage using non-uniform memory hierarchy for mobile intelligence, IEEE ISSCC Dig Tech Papers, 250
  3. bankman, 2018, An always-on $3.8~\mu\text{J}$ /86% CIFAR-10 mixed-signal binary CNN processor with all memory on chip in 28-nm CMOS, IEEE ISSCC Dig Tech Papers, 222
  4. horowitz, 2014, Computing’s energy problem (and what we can do about it), IEEE ISSCC Dig Tech Papers, 10

Cited by

Items 301–325 of 325

Metadata

Authors
Valavi, H., Ramadge, P. J., Nestler, E., & Verma, N.
Affiliations
  • Princeton University (US)
  • Analog Devices (United States) (US)
Venue
IEEE Journal of Solid-State Circuits
Published
2019-06
Volume
54
Issue
6
Pages
1789-1799
ISSN
  • 0018-9200 (ISSN-L)
  • 1558-173X
OpenAlex
W2920326572
Cited by
328
References
61
Citations by year
2026: 29 · 2025: 49 · 2024: 47 · 2023: 58 · 2022: 52 · 2021: 51 · 2020: 32 · 2019: 9
FWCI
12.90
Type
journal article
Language
en
Open access
closed
Locations
IEEE Journal of Solid-State Circuits (publishedVersion)
Funders
Analog Devices
Topics
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced Memory and Neural Computing
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Low-power high-performance VLSI design
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Analog and Mixed-Signal Circuit Design
Keywords
in-memory computing, CNN accelerator, matrix-vector multiplication, data movement reduction, metal-oxide-metal capacitors, binary neural networks, batch normalization, energy efficiency, 65 nm CMOS, convolutional layers

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.