An Oracle for Guiding Large-Scale Model/Hybrid Parallel Training of Convolutional Neural Networks

Abstract

Deep Neural Network (DNN) frameworks use distributed training to enable faster time to convergence and alleviate memory capacity limitations when training large models and/or using high dimension inputs. With the steady increase in datasets and model sizes, model/hybrid parallelism is deemed to have an important role in the future of distributed training of DNNs. We analyze the compute, communication, and memory requirements of Convolutional Neural Networks (CNNs) to understand the trade-offs between different parallelism approaches on performance and scalability. We leverage our model-driven analysis to be the basis for an oracle utility which can help in detecting the limitations and bottlenecks of different parallelism approaches at scale. We evaluate the oracle on six parallelization strategies, with four CNN models and multiple datasets (2D and 3D), on up to 1024 GPUs. The results demonstrate that the oracle has an average accuracy of about 86.74% when compared to empirical results, and as high as 97.57% for data parallelism. CCS CONCEPTS• Computing methodologies → Parallel computing methodologies; Distributed computing methodologies; Machine learning.

Comparative Study on Distributed Lightweight Deep Learning Models for Road Pothole Detection

“…This methodology assesses Oracle using six parallelization algorithms, four CNN models, and different datasets (2D and 3D) on up to 1024 GPUs. Compared to empirical results, the Oracle tool has an average accuracy of roughly 86.74% and data parallelism accuracy of up to 97.57% [25]. However, GPU processing performance and training throughput are severely limited because of the excessive memory consumption mentioned before.…”