GradiVeQ: Vector Quantization For Bandwidth-Efficient Gradient Aggregation In Distributed CNN Training

Authors:
Mingchao Yu University of Southern California
Zhifeng Lin University of Southern California
Krishna Narra University Of Southern California
Songze Li University of Southern California
Youjie Li UIUC
Nam Sung Kim University of Illinois at Urbana-Champaign
Alex Schwing University of Illinois at Urbana-Champaign
Murali Annavaram University of Southern California
Salman Avestimehr University of Southern California

Introduction:

Data parallelism can boost the training speed of convolutional neural networks (CNN), but could suffer from significant communication costs caused by gradient aggregation.In this paper, the authors empirically demonstrate the strong linear correlations between CNN gradients, and propose a gradient vector quantization technique, named GradiVeQ, to exploit these correlations through principal component analysis (PCA) for substantial gradient dimension reduction.

Abstract:

Data parallelism can boost the training speed of convolutional neural networks (CNN), but could suffer from significant communication costs caused by gradient aggregation. To alleviate this problem, several scalar quantization techniques have been developed to compress the gradients. But these techniques could perform poorly when used together with decentralized aggregation protocols like ring all-reduce (RAR), mainly due to their inability to directly aggregate compressed gradients. In this paper, we empirically demonstrate the strong linear correlations between CNN gradients, and propose a gradient vector quantization technique, named GradiVeQ, to exploit these correlations through principal component analysis (PCA) for substantial gradient dimension reduction. GradiveQ enables direct aggregation of compressed gradients, hence allows us to build a distributed learning system that parallelizes GradiveQ gradient compression and RAR communications. Extensive experiments on popular CNNs demonstrate that applying GradiveQ slashes the wall-clock gradient aggregation time of the original RAR by more than 5x without noticeable accuracy loss, and reduce the end-to-end training time by almost 50%. The results also show that \GradiveQ is compatible with scalar quantization techniques such as QSGD (Quantized SGD), and achieves a much higher speed-up gain under the same compression ratio.

You may want to know: