Conference abstractPublished 21 Aug 2018
Performance Modeling and Evaluation of Distributed Deep Learning Frameworks on GPUs
Hong Kong Bapist University, P.R.C
Abstract
Select any words in the title or abstract, or press this and choose a sentence.
Deep learning frameworks have been widely deployed on GPU servers for deep learning applications in both academia and industry. In training deep neural networks (DNNs), there are many standard processes or algorithms, such as convolution and stochastic gradient descent (SGD), but the running performance of different frameworks might be different even running the same deep model on the same GPU hardware. In this paper, we evaluate the running performance of four state-of-the-art distributed deep learning frameworks (i.e., Caffe-MPI, CNTK, MXNet and TensorFlow) over single-GPU, multi-GPU and multi-node environments. We first build performance models of standard processes in training DNNs with SGD, and then we benchmark the running performance of these frameworks with three popular convolutional neural networks (i.e., AlexNet, GoogleNet and ResNet-50), after that we analyze what factors that results in the performance gap among these four frameworks. Through both analytical and experimental analysis, we identify bottlenecks and overheads which could be further optimized. The main contribution is that the proposed performance models and the analysis provide further optimization directions in both algorithmic design and system configuration.
Ask about this article
Answers come only from this article's title, abstract and keywords, each sentence pointing to where it comes from. Your question is sent to OpenAI to write the answer. Scholars Central does not store it.
Follow
New articles from what you follow go into your email digest, once you agree to alert emails.
- SubjectEducation
- SubjectMathematics & Statistics
- BylineArticles whose byline prints “Xiowen Chu”
