The DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index.
A highly parameterizable framework for Conditional Restricted Boltzmann Machine based workloads accelerated with FPGAs and OpenCL
Open accessAbstract
Conditional Restricted Boltzmann Machine (CRBM) is a promising candidate for a multidimensional system modeling that can learn a probability distribution over a set of data. It is a specific type of an artificial neural network with one input (visible) and one output (hidden) layer. Recently published works demonstrate that CRBM is a suitable mechanism for modeling multidimensional time series such as human motion, workload characterization, city traffic analysis. The process of learning and inference of these systems relies on linear algebra functions like matrix–matrix multiplication, and for higher data sets, they are very compute-intensive. In this paper, we present a configurable framework for CRBM based workloads for arbitrary large models. We show how to accelerate the learning process of CRBM with FPGAs and OpenCL, and we conduct an extensive scalability study for different model sizes and system configurations. We show significant improvement in performance/Watt for large models and batch sizes (from 1.51x up to 5.71x depending on the host configuration) when we use FPGA and OpenCL for the acceleration, and limited benefits for small models comparing to the state-of-the-art CPU solution.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Their work showed a 15% execution time improvement by using host pipes to reduce the data transfer overhead between CPU and FPGA [15].”
From Embedded Intelligence on FPGA: Survey, Applications and Challenges (2021) · cites this paper as [15]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“By using FPGA on the CRBM, it is able to save on power consumption and increase performance by approximately five times higher than CPU solutions.”
From Embedded Intelligence on FPGA: Survey, Applications and Challenges (2021) · cites this paper as [15]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
An OpenCL™ Deep Learning Accelerator on Arria 10
Aydonat, U., O’Connell, S., Capalija, D., Ling, A. C., & Chiu, G. R. (2017) · ACM/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA) · doi:10.1145/3020078.3021738 · Cited by 229
Improving the Performance of OpenCL-based FPGA Accelerator for Convolutional Neural Network
Zhang, J., & Li, J. (2017) · ACM/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA) · doi:10.1145/3020078.3021698 · Cited by 212
Demystifying Parallel and Distributed Deep Learning: An In-Depth Concurrency AnalysisPreprint
Ben‐Nun, T., & Hoefler, T. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1802.09941 · Cited by 204
A Survey of FPGA-Based Neural Network AcceleratorPreprint
Guo, K., Zeng, S., Yu, J., Wang, Y., & Yang, H. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1712.08934 · Cited by 147
Conditional Restricted Boltzmann Machines for Structured Output PredictionPreprint
Mnih, V., Larochelle, H., & Hinton, G. E. (2012) · arXiv (Cornell University) · doi:10.48550/arxiv.1202.3748 · Cited by 123
Two Distributed-State Models For Generating High-Dimensional Time Series
Taylor, G. W., Hinton, G. E., & Roweis, S. T. (2011) · doi:10.5555/1953048.2021035 · Cited by 114
Understanding Performance Differences of FPGAs and GPUs
Cong, J., Fang, Z., Lo, M., et al. (2018) · IEEE International Symposium on Field-Programmable Custom Computing Machines (FCCM) · doi:10.1109/fccm.2018.00023 · Cited by 106
Proceedings of the 2018 ACM/SIGDA International Symposium on Field-Programmable Gate Arrays
Ng, H.-C., Liu, S., & Luk, W. (2018) · doi:10.1145/3174243 · Cited by 102
Are FPGAs Suitable for Edge Computing?
Biookaghazadeh, S., Ren, F., & Zhao, M. (2018) · arXiv · doi:10.48550/arxiv.1804.06404 · Cited by 83
A Customizable Matrix Multiplication Framework for the Intel HARPv2 Xeon+FPGA Platform
Moss, D. J. M., Krishnan, S., Nurvitadhi, E., Ratuszniak, P., Johnson, C., Sim, J., Mishra, A., Marr, D., Subhaschandra, S., & Leong, P. H. W. (2018) · ACM/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA) · doi:10.1145/3174243.3174258 · Cited by 81
A highly scalable Restricted Boltzmann Machine FPGA implementation
Kim, S. K., McAfee, L., McMahon, P. L., & Olukotun, K. (2009) · International Conference on Field Programmable Logic and Applications (FPL) · doi:10.1109/fpl.2009.5272262 · Cited by 80
DeepX: Deep Learning Accelerator for Restricted Boltzmann Machine Artificial Neural Networks
Kim, L.-W. (2017) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2017.2665555 · Cited by 71
A high-performance FPGA architecture for restricted boltzmann machines
Ly, D. L., & Chow, P. (2009) · ACM/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA) · doi:10.1145/1508128.1508140 · Cited by 34
An FPGA implementation of a Restricted Boltzmann Machine classifier using stochastic bit streams
Li, B., Najafi, M. H., & Lilja, D. J. (2015) · IEEE International Conference on Application-specific Systems, Architectures and Processors (ASAP) · doi:10.1109/asap.2015.7245709 · Cited by 33
Considerations in using OpenCL on GPUs and FPGAs for throughput-oriented genomics workloads
Cadenelli, N., Jakšić, Z., Polo, J., et al. (2018) · Future Generation Computer Systems · doi:10.1016/j.future.2018.11.028 · Cited by 30
Self-citationAutomatic Generation of Workload Profiles Using Unsupervised Learning Pipelines
Buchaca, D., Berral, J. L., & Carrera, D. (2017) · IEEE Transactions on Network and Service Management · doi:10.1109/tnsm.2017.2786047 · Cited by 19
FPGA Implementation of a Scalable and Highly Parallel Architecture for Restricted Boltzmann Machines
Ueyoshi, K., Marukame, T., Asai, T., Motomura, M., & Schmid, A. (2016) · Circuits and Systems · doi:10.4236/cs.2016.79185 · Cited by 14
Customizable FPGA OpenCL matrix multiply design template for deep neural networks
Yinger, J., Nurvitadhi, E., Capalija, D., Ling, A. C., Marr, D., Krishnan, S., Moss, D. J. M., & Subhaschandra, S. (2017) · International Conference on Field-Programmable Technology (FPT) · doi:10.1109/fpt.2017.8280155 · Cited by 13
Moss Duncan, J. M., Krishnan, S., Nurvitadhi, E., et al. (2018) · ACM Proceedings · Cited by 9
Evaluation of ‘OpenCL for FPGA’ for Data Acquisition and Acceleration in High Energy Physics
Sridharan, S. (2015) · Journal of Physics Conference Series · doi:10.1088/1742-6596/664/9/092023 · Cited by 8
Research works printed but not linked: 3
- D. Castells-Rufas, J. Carrabina, Opencl-based FPGA accelerator for disparity map generation with stereoscopic event cameras, CoRR abs/1903.03509. arXiv:1903.03509. URL http://arxiv.org/abs/1903.03509. · arXiv
- C. Lo, An fpga implementation of large restricted boltzmann machines, in: Master Thesis, 2010.
- T. Xia, Fpga implementation of a restricted boltzmann machine for handwriting recognition, in: Master Thesis, 2015.
Cited by
Sorted by citations
Embedded Intelligence on FPGA: Survey, Applications and Challenges
Seng, K. P., Lee, P. J., & Ang, L. M. (2021) · Electronics · doi:10.3390/electronics10080895 · Cited by 105
Metadata
- Authors
- Jakšić, Z., Cadenelli, N., Prats, D. B., Polo, J., Berral Garcia, J. L., & Perez, D. C.
- Affiliations
- Barcelona Supercomputing Center (ES)
- Corresponding author
- Jakšić, Zoran
- Venue
- Future Generation Computer Systems
- Publisher
- Elsevier BV
- Published
- 2020-03
- Volume
- 104
- Pages
- 201-211
- ISSN
- 0167-739X (ISSN-L)
- OpenAlex
- W2987836951
- Cited by
- 10
- References
- 26
- Citations by year
- 2025: 3 · 2024: 2 · 2022: 3 · 2021: 1 · 2020: 1
- FWCI
- 1.27
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- Future Generation Computer Systems (publishedVersion, cc-by)
- LA Referencia (Red Federada de Repositorios Institucionales de Publicaciones Científicas) (acceptedVersion, cc-by-nc-nd)
- UPCommons institutional repository (Universitat Politècnica de Catalunya) (submittedVersion, cc-by-nc-nd)
- Funders
- Intel Corporation
- Generalitat de Catalunya
- European Research Council
- Barcelona Supercomputing Center
- Topics
- Physical Sciences › Computer Science › Hardware and Architecture › Embedded Systems Design Techniques
- Physical Sciences › Computer Science › Hardware and Architecture › Parallel Computing and Optimization Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Keywords
- conditional restricted Boltzmann machine, FPGA acceleration, OpenCL, workload characterization, energy efficiency, performance per watt, matrix multiplication
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
