Abstract
Deep learning and convolutional neural network (CNN) have been shown to solve image classification problems fast and with high accuracy. However, these algorithms tend to be very computationally intensive and resource hungry, hence making them difficult to use on embedded devices. Towards this end, we need system-level models for analysis and simulation. In this report, we describe a newly designed untimed SystemC model of GoogLeNet, a state-of-the-art deep CNN using OpenCV library. The SystemC model is automatically created from a Caffe model using a generator tool. We successfully validate the functionality of the model using Accellera SystemC 2.3.1 simulator. Then, we use RISC (Recoding Infrastructure for SystemC) to speed up the simulation by exploiting thread-level parallelism and report extensive experimental results.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
OpenCV has built with -O0 flag meaning (almost) no compiler optimizations.
References
Gerstlauer, A., Dömer, R., Peng, J., Gajski, D.D.: System Design: A Practical Guide with SpecC. Kluwer (2001)
Grötker, T., Liao, S., Martin, G., Swan, S.: System Design with SystemC. Kluwer (2002)
Huang, G., Liu, Z., van der Maaten, L., Weinberger, K.Q.: Densely connected convolutional networks. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017, pp. 2261–2269. IEEE Computer Society (2017)
Jia, Y., et al.: Caffe: convolutional architecture for fast feature embedding. arXiv preprint arXiv:1408.5093 (2014)
Krizhevsky, A., Sutskever, I., Hinton, G.E.: ImageNet classification with deep convolutional neural networks. In: NIPS (2012)
Larsson, G., Maire, M., Shakhnarovich, G.: FractalNet: ultra-deep neural networks without residuals. CoRR abs/1605.07648 (2016)
Le Cun, Y., et al.: Handwritten digit recognition: applications of neural network chips and automatic learning. Commun. Mag. 27(11), 41–46 (1989)
LeCun, Y., Haffner, P., Bottou, L., Bengio, Y.: Object recognition with gradient-based learning. In: Shape, Contour and Grouping in Computer Vision. LNCS, vol. 1681, pp. 319–345. Springer, Heidelberg (1999). https://doi.org/10.1007/3-540-46805-6_19
Liu, G., Schmidt, T., Cheng, Z., Mendoza, D., Dömer, R.: RISC Compiler and Simulator, Release V0.5.0: Out-of-Order Parallel Simulatable SystemC Subset. Technical report CECS-TR-18-03, CECpS, UCI (2018)
OpenCV Tutorials, Load Caffe framework models. https://docs.opencv.org/3.4/d5/de7/tutorial_dnn_googlenet.html. Accessed 11 May 2019
Sze, V., Chen, Y., Yang, T., Emer, J.S.: Efficient processing of deep neural networks: a tutorial and survey. Proc. IEEE 105(12), 2295–2329 (2017)
Szegedy, C., et al.: Going deeper with convolutions. In: IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2015, pp. 1–9 (2015)
Xie, S., Girshick, R.B., Dollár, P., Tu, Z., He, K.: Aggregated residual transformations for deep neural networks. CoRR abs/1611.05431 (2016)
Author information
Authors and Affiliations
Corresponding authors
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 IFIP International Federation for Information Processing
About this paper
Cite this paper
Arasteh, E.M., Dömer, R. (2023). An Untimed SystemC Model of GoogLeNet. In: Wehrmeister, M.A., Kreutz, M., Götz, M., Henkler, S., Pimentel, A.D., Rettberg, A. (eds) Analysis, Estimations, and Applications of Embedded Systems. IESS 2019. IFIP Advances in Information and Communication Technology, vol 576. Springer, Cham. https://doi.org/10.1007/978-3-031-26500-6_10
Download citation
DOI: https://doi.org/10.1007/978-3-031-26500-6_10
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-26499-3
Online ISBN: 978-3-031-26500-6
eBook Packages: Computer ScienceComputer Science (R0)