The DOI matches its Crossref record; 79 of the 79 references it cites were found in a registry or an index.
A Heterogeneous and Programmable Compute-In-Memory Accelerator Architecture for Analog-AI Using Dense 2-D Mesh
Abstract
We introduce a highly heterogeneous and programmable compute-in-memory (CIM) accelerator architecture for deep neural network (DNN) inference. This architecture combines spatially distributed CIM memory array “tiles” for weight-stationary, energy-efficient multiply–accumulate (MAC) operations, together with heterogeneous special-function compute cores for auxiliary digital computation. Massively parallel vectors of neuron activation data are exchanged over short distances using a dense and efficient circuit-switched 2-D mesh, offering full end-to-end support for a wide range of DNN workloads, including CNNs, long-short-term-memory (LSTM), and transformers. We discuss the design of the “analog fabric”—the 2-D grid of tiles and compute cores interconnected by the 2-D mesh—and address the efficiency in both mapping of DNNs onto the hardware and in pipelining of various DNN workloads across a range of batch sizes. We show, for the first time, system-level assessments using projected component parameters for a realistic “analog AI” system, based on dense crossbar arrays of low-power nonvolatile analog memory elements, while incorporating a single common analog fabric design that can scale to large networks by introducing data transport between multiple analog AI chips. Our performance estimates for several networks, including large LSTM and bidirectional encoder representations from transformers (BERT), show highly competitive throughput while offering$40\times $–$140\times $higher energy efficiency than NVIDIA A100—thus illustrating the strong promise of analog AI and the proposed architecture for DNN inference applications.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
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?
In the citing paper
“In terms of AIMC workload execution latency and system mapping51, CNNs are already less well-suited for resistive crossbar arrays due to the uneven temporal reuse between layers and spatial under-utilization of the large analog tiles by the small kernel matrices (see Table 1), although some optimization and mapping tricks52 are available.”
From Hardware-aware training for large-scale and diverse deep learning inference workloads using in-memory computing-based accelerators (2023) · cites this paper as [51]
In the citing paper
“Interestingly, RNNs—already well-suited for AIMC given their large, dense MVMs51—also seem to be the most robust to the finite signal-to-noise ratio (SNR) of AIMC hardware.”
From Hardware-aware training for large-scale and diverse deep learning inference workloads using in-memory computing-based accelerators (2023) · cites this paper as [51]
In the citing paper
“The advantages lie with the mature design and fabrication processes, and the shortcomings involve limitations to traditional computer architectures, such as data movement bottleneck [239, 240], rather than, e.g., synaptic time-dependent neuromorphic operations.”
From Brain-inspired computing systems: a systematic literature review (2024) · cites this paper as [239]
In the citing paper
“A highly heterogeneous and programmable accelerator architecture for analog AI has been introduced20 for which system-level performance assessments have predicted energy efficiencies 40–140 times higher than those of cutting-edge graphics processing units.”
From An analog-AI chip for energy-efficient speech recognition and transcription (2023) · cites this paper as [20]
In the citing paper
“The first is the use of a dense and efficient circuit-switched 2D mesh to exchange massively parallel vectors of neuron-activation data over short distances.”
From An analog-AI chip for energy-efficient speech recognition and transcription (2023) · cites this paper as [20]
3 more claims about this paper
Create a free account to see every claim other papers make about this one.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 1×
“…transducers (RNN-T) model used for speech-to-text transcription [49], two bidirectional encoder representations from transformers (BERT) [50] models used in natural language processing (NLP) applications, and two advanced CNN models (ResNet-18 and -50 [51]) used for image processing applications.”
MethodVaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404 · cited 2×
“As discussed in Section II-A, auxiliary Ops are a small but critical part of DNN workloads, especially for LSTM [54] and Transformers [55].”
BackgroundHochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 1×
“As discussed in Section II-A, auxiliary Ops are a small but critical part of DNN workloads, especially for LSTM [54] and Transformers [55].”
BackgroundBERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173 · cited 1×
“…network transducers (RNN-T) model used for speech-to-text transcription [49], two bidirectional encoder representations from transformers (BERT) [50] models used in natural language processing (NLP) applications, and two advanced CNN models (ResNet-18 and -50 [51]) used for image processing…”
MethodBERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325 · cited 1×
“…network transducers (RNN-T) model used for speech-to-text transcription [49], two bidirectional encoder representations from transformers (BERT) [50] models used in natural language processing (NLP) applications, and two advanced CNN models (ResNet-18 and -50 [51]) used for image processing…”
MethodLanguage Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372 · cited 1×
“Better hardware enabled the training of very large networks with enormous datasets [1], [2], as well as rapid inference of the resulting large and thus highly capable DNN models [2].”
BackgroundKaugeranna, R., Kaugeranna, E., & Aion, (. S. 4. 6. (2023) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.4230/lipics.giscience.2023.43 · Cited by 13,690
Han, S., Mao, H., & Dally, W. J. (2015) · arXiv · doi:10.48550/arxiv.1510.00149 · Cited by 4,662 · cited 2×
“…features designed specifically for DNN training and inference [3], to specialized digital CMOS accelerators [4], [5], [6], [7] incorporating reduced precision [8], sparsity [9], dataflow architectures [10], [11], hardware–software optimization [12], [13], and very-large-area accelerator chips [14].”
BackgroundMethodIn-Datacenter Performance Analysis of a Tensor Processing Unit
Jouppi, N. P., Young, C., Patil, N., Patterson, D., Agrawal, G., Bajwa, R., Bates, S., Bhatia, S., Boden, N., Borchers, A., et al. (2017) · Proceedings of the 44th Annual International Symposium on Computer Architecture · doi:10.1145/3079856.3080246 · Cited by 4,537 · cited 2×
“Those based on more conventional computing techniques range from low-power ASICs and reconfig-urable cores [6], [67], [68], [69] to large-scale AI systems [10], [70].”
BackgroundMethodValidity of the single processor approach to achieving large scale computing capabilities
Amdahl, G. M. (1967) · Proceedings of the April 18-20, 1967, spring joint computer conference on - AFIPS '67 (Spring) · doi:10.1145/1465482.1465560 · Cited by 4,339 · cited 1×
“We discuss this phenomenon—analogous to “Amdahl’s law” for parallel computing [52]—below.”
Research works printed but not linked: 8
- Amodei, 2018, AI and compute
- 2022, Nvidia Data Center Deep Learning Product Performance
- Sarma, 2019, Compute and redundancy solution for Tesla’s full self driving computer
- Lie, 2019, Wafer-scale deep learning
- Fick, 2018, Analog computation in flash memory for datacenter-scale AI inference in a small chip
- Strachan, 2015, The dot-product engine (DPE): Exploring high efficiency analog multiplication with memristor arrays
- 2022, MLPerf Inference Benchmark Suite
- Yang, Sparse ReRAM engine: Joint exploration of activation and weight sparsity in compressed neural networks, 236
Cited by
Hardware implementation of memristor-based artificial neural networks
Aguirre, F., Sebastian, A., Le Gallo, M., Song, W., Wang, T., Yang, J. J., Lu, W., Chang, M.-F., Ielmini, D., Yang, Y., et al. (2024) · Nature Communications · doi:10.1038/s41467-024-45670-9 · Cited by 564
Self-citationRecent Advances and Future Prospects for Memristive Materials, Devices, and Systems
Song, M.-K., Kang, J.-H., Zhang, X., Ji, W., Ascoli, A., Messaris, I., Demirkol, A. S., Dong, B., Aggarwal, S., Wan, W., et al. (2023) · ACS Nano · doi:10.1021/acsnano.3c03505 · Cited by 300
Self-citationLe Gallo, M., Khaddam-Aljameh, R., Stanisavljevic, M., Vasilopoulos, A., Kersting, B., Dazzi, M., Karunaratne, G., Brändli, M., Singh, A., Müller, S. M., et al. (2023) · Nature Electronics · doi:10.1038/s41928-023-01010-1 · Cited by 310
Self-citationAn analog-AI chip for energy-efficient speech recognition and transcription
Ambrogio, S., Narayanan, P., Okazaki, A., Fasoli, A., Mackin, C., Hosokawa, K., Nomura, A., Yasuda, T., Chen, A., Friz, A., et al. (2023) · Nature · doi:10.1038/s41586-023-06337-5 · Cited by 269
“A highly heterogeneous and programmable accelerator architecture for analog AI has been introduced 20 for which system-level performance assessments have predicted energy efficiencies 40–140 times higher than those of cutting-edge graphics processing units.”
BackgroundMethodInfluentialSelf-citationRasch, M. J., Mackin, C., Le Gallo, M., Chen, A., Fasoli, A., Odermatt, F., Li, N., Nandakumar, S. R., Narayanan, P., Tsai, H., et al. (2023) · Nature Communications · doi:10.1038/s41467-023-40770-4 · Cited by 213
Self-citation
Metadata
- Authors
- Jain, S., Tsai, H., Chen, C.-T., Muralidhar, R., Boybat, I., Frank, M. M., Woźniak, S., Stanisavljevic, M., Adusumilli, P., Narayanan, P., Hosokawa, K., Ishii, M., Kumar, A., Narayanan, V., & Burr, G. W.
- Affiliations
- IBM (United States) (US)
- IBM Research–Almaden, San Jose, CA, USA
- IBM Research - Zurich (CH)
- IBM Research Zurich, Rüschlikon, Switzerland
- IBM Research - Almaden (US)
- IBM Research - Tokyo (JP)
- ORCID
- Venue
- IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Published
- 2023-01
- Volume
- 31
- Issue
- 1
- Pages
- 114-127
- ISSN
- 1063-8210 (ISSN-L)
- 1557-9999
- OpenAlex
- W4313251083
- Cited by
- 76
- References
- 78
- Citations by year
- 2026: 11 · 2025: 34 · 2024: 19 · 2023: 12
- FWCI
- 4.45
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- IEEE Transactions on Very Large Scale Integration (VLSI) Systems (publishedVersion)
- Topics
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced Memory and Neural Computing
- Physical Sciences › Computer Science › Hardware and Architecture › Parallel Computing and Optimization Techniques
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Analog and Mixed-Signal Circuit Design
- Keywords
- in-memory computing, multiply-accumulate operations, crossbar arrays, convolutional neural networks, LSTM, BERT, energy efficiency, pipelining
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: A Heterogeneous and Programmable Compute-In-Memory Accelerator Architecture for Analog-AI Using Dense 2-D Mesh. https://app.cochat.ai/papers/10.1109/tvlsi.2022.3221390
- The paper
- Jain, S., Tsai, H., Chen, C.-T., Muralidhar, R., Boybat, I., Frank, M. M., Woźniak, S., Stanisavljevic, M., Adusumilli, P., Narayanan, P., Hosokawa, K., Ishii, M., Kumar, A., Narayanan, V., & Burr, G. W. (2023). A Heterogeneous and Programmable Compute-In-Memory Accelerator Architecture for Analog-AI Using Dense 2-D Mesh. IEEE Transactions on Very Large Scale Integration (VLSI) Systems. https://doi.org/10.1109/tvlsi.2022.3221390
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.