PUMA: A Programmable Ultra-efficient Memristor-based Accelerator for Machine Learning Inference
The DOI matches its Crossref record; 118 of the 121 references it cites were found in a registry or an index.
Abstract
Memristor crossbars are circuits capable of performing analog matrix-vector multiplications, overcoming the fundamental energy efficiency limitations of digital logic. They have been shown to be effective in special-purpose accelerators for a limited set of neural network applications. We present the Programmable Ultra-efficient Memristor-based Accelerator (PUMA) which enhances memristor crossbars with general purpose execution units to enable the acceleration of a wide variety of Machine Learning (ML) inference workloads. PUMA's microarchitecture techniques exposed through a specialized Instruction Set Architecture (ISA) retain the efficiency of in-memory computing and analog circuitry, without compromising programmability. We also present the PUMA compiler which translates high-level code to PUMA ISA. The compiler partitions the computational graph and optimizes instruction scheduling and register allocation to generate code for large and complex workloads to run on thousands of spatial cores. We have developed a detailed architecture simulator that incorporates the functionality, timing, and power models of PUMA's components to evaluate performance and energy consumption. A PUMA accelerator running at 1 GHz can reach area and power efficiency of 577 GOPS/s/mm 2 and 837~GOPS/s/W, respectively. Our evaluation of diverse ML applications from image recognition, machine translation, and language modelling (5M-800M synapses) shows that PUMA achieves up to 2,446× energy and 66× latency improvement for inference compared to state-of-the-art GPUs. Compared to an application-specific memristor-based accelerator, PUMA incurs small energy overheads at similar inference latency and added programmability.
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
“ADC can consume up to 70–90% of the on-chip area of the crossbar-based computation unit, including memristive crossbar and peripheral circuits, and up to 80-88% of energy55,154,155. In summary, ADCs are commonly the largest and most power-hungry circuit block in a memristive neural network55,156.”
From Hardware implementation of memristor-based artificial neural networks (2024) · cites this paper as [154]
In the citing paper
“ADC can consume up to 70–90% of the on-chip area of the crossbar-based computation unit, including memristive crossbar and peripheral circuits, and up to 80-88% of energy55,154,155.”
From Hardware implementation of memristor-based artificial neural networks (2024) · cites this paper as [154]
Citations
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318
GAN(Generative Adversarial Nets)
(2017) · Journal of Japan Society for Fuzzy Theory and Intelligent Informatics · doi:10.3156/jsoft.29.5_177_2 · Cited by 19,270
Applied Linear Statistical Models
Ziegel, E. R., Neter, J., Wasserman, W., & Kutner, M. (1992) · Technometrics · doi:10.2307/1269588 · Cited by 9,907
Applied Linear Statistical Models.
Lorenz, F., Neter, J., Wasserman, W., & Kutner, M. H. (1986) · Journal of the American Statistical Association · doi:10.2307/2289108 · Cited by 5,565
Recurrent neural network based language model
Mikolov, T., Karafiát, M., Burget, L., Černocký, J., & Khudanpur, S. (2010) · Interspeech 2010 · doi:10.21437/interspeech.2010-343 · Cited by 5,390 · cited 1×
“Recurrent Neural Networks (RNNs) [75] used for sequence processing compute weighted-sums on input and previous state.”
MethodRedox‐Based Resistive Switching Memories – Nanoionic Mechanisms, Prospects, and Challenges
Waser, R., Dittmann, R., Staikov, G., & Szot, K. (2009) · Advanced Materials · doi:10.1002/adma.200900375 · Cited by 5,080 · cited 1×
“Each crosspoint in the crossbar stores a multi-bit value in one memristor device, which enables high storage density [112].”
BackgroundIn-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
Efficient Processing of Deep Neural Networks: A Tutorial and Survey
Sze, V., Chen, Y.-H., Yang, T.-J., & Emer, J. S. (2017) · Proceedings of the IEEE · doi:10.1109/jproc.2017.2761740 · Cited by 4,215 · cited 1×
“[103] provide a thorough survey of deep learning accelerators.”
BackgroundResnick, P., & Varian, H. R. (1997) · Communications of the ACM · doi:10.1145/245108.245121 · Cited by 3,772
Research works printed but not linked: 16
- Alan Agresti. 2002. Logistic regression .Wiley Online Library.
- 10.5555/3195638.3195664 · doi:10.5555/3195638.3195664
- B. Feinberg, S. Wang, and E. Ipek. 2018b. Making Memristive Neural Network Accelerators Reliable. In 2018 IEEE International Symposium on High Performance Computer Architecture (HPCA) .
- Ian Goodfellow Jean Pouget-Abadie Mehdi Mirza Bing Xu David Warde-Farley Sherjil Ozair Aaron Courville and Yoshua Bengio. 2014. Generative adversarial nets. In Advances in neural information processing systems. 2672--2680. · doi:10.5555/2969033.2969125
- Kaiyuan Guo, Shulin Zeng, Jincheng Yu, Yu Wang, and Huazhong Yang. 2017b. A Survey of FPGA Based Neural Network Accelerator. arXiv preprint arXiv:1712.08934 (2017). · arXiv
- 10.5555/3045118.3045303 · doi:10.5555/3045118.3045303
- 10.1145/3173162.3173205 · DOI
- Nan Jiang, George Michelogiannakis, Daniel Becker, Brian Towles, and William J. Dally. 2013. BookSim 2.0 User's Guide .
- 10.5555/3195638.3195661 · doi:10.5555/3195638.3195661
- 10.5555/3195638.3195693 · doi:10.5555/3195638.3195693
- redo logging for persistent memory systems. In High Performance Computer Architecture (HPCA), 2018 IEEE International Symposium on. IEEE, 336--349.
- 10.5555/3195638.3195659 · doi:10.5555/3195638.3195659
- Patrick M Sheridan, Fuxi Cai, Chao Du, Wen Ma, Zhengya Zhang, and Wei D Lu. 2017. Sparse coding with memristor networks. Nature nanotechnology (2017).
- 10.5555/3195638.3195696 · doi:10.5555/3195638.3195696
- Richard Yu. 2017. Panasonic and UMC Partner for 40nm ReRAM Process Platform. UMC Press Release (Feb 2017).
- 10.5555/3195638.3195662 · doi:10.5555/3195638.3195662
Cited by
Memory devices and applications for in-memory computing
Sebastian, A., Le Gallo, M., Khaddam-Aljameh, R., & Eleftheriou, E. (2020) · Nature Nanotechnology · doi:10.1038/s41565-020-0655-z · Cited by 2,289
Model Compression and Hardware Acceleration for Neural Networks: A Comprehensive Survey
Deng, B. L., Li, G., Han, S., Shi, L., & Xie, Y. (2020) · Proceedings of the IEEE · doi:10.1109/jproc.2020.2976475 · Cited by 996
“Besides the MVM operation, some other operations such as scalar and vector operations should be additionally supported by PEs since they are necessary for some neural network models such as RNNs but cannot be handled by the memristor array efficiently [258].”
MethodInfluentialA compute-in-memory chip based on resistive random-access memory
Wan, W., Kubendran, R., Schaefer, C., Eryilmaz, S. B., Zhang, W., Wu, D., Deiss, S., Raina, P., Qian, H., Gao, B., et al. (2022) · Nature · doi:10.1038/s41586-022-04992-8 · Cited by 943
Neuro-inspired computing chips
Zhang, W., Gao, B., Tang, J., Yao, P., Yu, S., Chang, M.-F., Yoo, H.-J., Qian, H., & Wu, H. (2020) · Nature Electronics · doi:10.1038/s41928-020-0435-7 · Cited by 892
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
“Moreover, ADC can consume up to 70 – 90% of the on-chip area of the crossbar-based computation unit, including memristive crossbar and peripheral circuits, and up to 80-88% of energy 55,154,155 .”
Metadata
- Authors
- Ankit, A., Hajj, I. E., Chalamalasetti, S. R., Ndu, G., Foltin, M., Williams, R. S., Faraboschi, P., Hwu, W.-M. W., Strachan, J. P., Roy, K., & Milojicic, D. S.
- Affiliations
- Purdue University West Lafayette (US)
- University of Illinois Urbana-Champaign (US)
- Hewlett Packard Enterprise (United States) (US)
- Hewlett-Packard (United Kingdom) (GB)
- Hewlett Packard Enterprise, Fort Collins, CO, USA
- ORCID
- Venue
- Proceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Systems
- Publisher
- ACM
- Published
- 2019-04-04
- Pages
- 715-731
- OpenAlex
- W2913104037
- Cited by
- 424
- References
- 121
- Citations by year
- 2026: 32 · 2025: 67 · 2024: 49 · 2023: 64 · 2022: 68 · 2021: 65 · 2020: 60 · 2019: 16
- FWCI
- 83.15
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) (publishedVersion)
- Funders
- U.S. Department of Energy (DE-SC0012199)
- Intelligence Advanced Research Projects Activity (2017-1701300000)
- Center for Brain-inspired Computing (C-BRIC), one of six centers in JUMP, a DARPA sponsored Semiconductor Research Corporation (SRC) program
- 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 › Low-power high-performance VLSI design
- Keywords
- memristor crossbar, in-memory computing, instruction set architecture, instruction scheduling, register allocation, energy efficiency, image recognition, machine translation, language modeling
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: PUMA: A Programmable Ultra-efficient Memristor-based Accelerator for Machine Learning Inference. https://app.cochat.ai/papers/10.1145/3297858.3304049
- The paper
- Ankit, A., Hajj, I. E., Chalamalasetti, S. R., Ndu, G., Foltin, M., Williams, R. S., Faraboschi, P., Hwu, W.-M. W., Strachan, J. P., Roy, K., & Milojicic, D. S. (2019). PUMA: A Programmable Ultra-efficient Memristor-based Accelerator for Machine Learning Inference. Proceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Systems. https://doi.org/10.1145/3297858.3304049
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.