MergeComp: A Compression Scheduler for Scalable Communication-Efficient Distributed Training
arXiv:2103.15195
Abstract
Large-scale distributed training is increasingly becoming communication bound. Many gradient compression algorithms have been proposed to reduce the communication overhead and improve scalability. However, it has been observed that in some cases gradient compression may even harm the performance of distributed training. In this paper, we propose MergeComp, a compression scheduler to optimize the scalability of communication-efficient distributed training. It automatically schedules the compression operations to optimize the performance of compression algorithms without the knowledge of model architectures or system parameters. We have applied MergeComp to nine popular compression algorithms. Our evaluations show that MergeComp can improve the performance of compression algorithms by up to 3.83x without losing accuracy. It can even achieve a scaling factor of distributed training up to 99% over high-speed networks.
arXiv comment8 papes
References in corpus (14)
- PyTorch: An Imperative Style, High-Performance Deep Learning Library
- MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems
- QSGD: Communication-Efficient SGD via Gradient Quantization and Encoding
- Horovod: fast and easy distributed deep learning in TensorFlow
- Sparsified SGD with Memory
- Don't Use Large Mini-Batches, Use Local SGD
- Local SGD Converges Fast and Communicates Little
- PowerSGD: Practical Low-Rank Gradient Compression for Distributed Optimization
- PyTorch Distributed: Experiences on Accelerating Data Parallel Training
- Scaling Distributed Machine Learning with In-Network Aggregation
- 8-Bit Approximations for Parallelism in Deep Learning
- AdaComp : Adaptive Residual Gradient Compression for Data-Parallel Distributed Training
- Variance-based Gradient Compression for Efficient Distributed Deep Learning
- Error Compensated Quantized SGD and its Applications to Large-scale Distributed Optimization