Abstract
Convolutional Neural networks (CNN) have succeeded great impact in various tasks of machine learning. Training CNN model is a computationally intensive task. Scalability and performance of CNN with GPU is demonstrated in this study using CUDA (Compute Unified Device Architecture) framework. We evaluated performance characteristics using our own designed CNN model. The model is configured for Facial expression Recognition classification task. The novelty of this effort is to demonstrate performance acceleration in scalable CNN. The parallel tasks using the hardware feature of general-purpose- computing on graphics processing unit (GPGPU) has been shown to be appropriate to be applied to CNN. We implemented at multi-node distributed training; which allows us to efficiently parallelize deep networks across multiple servers, in order to minimize the time to train. The experiment result shows that the proposed experiment in this study gained over 7 times speedup.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsReferences
LeCun, Y., Bengio, Y.: Pattern recognition and Neural Networks. In: Arbib, M.A. (ed.) Handbook of Brain Theory and Neural Networks. MIT Press (1995)
Chellapilla, K., Puri, S., Simard, P.: High performance convolutional neural networks for document processing. In: Tenth International Workshop on Frontiers in Handwriting Recognition, Université de Rennes 1, Oct 2006, La Baule (2006)
Li, X., et al.: Performance analysis of GPU-based convolutional neural networks. In: 45th International Conference on Parallel Processing (2016)
Arriaga, O., Valdenegro-Toro, M., Plöger, P.: Real-time convolutional neural networks for emotion and gender classification. arXiv preprint arXiv:1710.07557 (2017)
Guo, Y., Tao, D., Xiang, H.: Deep neural networks with relativity learning for facial expression recognition. In: IEEE International Conference on Multimedia & Expo Workshops (ICMEW) (2016)
Huang, Y., Li, K., Wang, G., Cao, M., Li, P., Zhang, Y.: Recognition of convolutional neural network based on CUDA technology (2015). Available https://arxiv.org/abs/1506.00074 [Online]
Adam, B., Zaman, F.H.K., Yassin, I.M., Abidin, H.Z.: Faster R-CNN Implementation using CUDA Architecture in GeForce GTX 10 Series (2016)
Choi, S., Lee, K.: A CUDA-based implementation of convolutional neural network. In: 4th International Conference on Computer Applications and Information Processing Technology (CAIPT), Kuta Bali (2017)
Stephan, P., Daniel, S., Klaus, K.: Performance and scalability of GPU based convolutional neural networks. IEEE (2010)
Corrado, G., Monga, R., Chen, K., Devin, M., Senior, A., Tucker, P.: Large scale distributed deep networks. In: NIPS (2012)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Chavan, U., Kulkarni, D. (2020). Performance Analysis of Parallel and Scalable GPU Based Convolutional Neural Network. In: Iyer, B., Deshpande, P., Sharma, S., Shiurkar, U. (eds) Computing in Engineering and Technology. Advances in Intelligent Systems and Computing, vol 1025. Springer, Singapore. https://doi.org/10.1007/978-981-32-9515-5_45
Download citation
DOI: https://doi.org/10.1007/978-981-32-9515-5_45
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-32-9514-8
Online ISBN: 978-981-32-9515-5
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)