PyTorch Distributed: Experiences on Accelerating Data Parallel Training
PyTorch's distributed data parallel module uses gradient bucketing, communication-computation overlap, and synchronization skipping to achieve near-linear scalability on 256 GPUs.
Published Jun 28, 2020111 citations▲ 13 on Hugging FaceCode ★ 103,810arXiv ↗
Only vote on papers you've read. Sign in with GitHub to vote.
PyTorch's distributed data parallel delivers credible near-linear scaling to 256 GPUs through gradient bucketing and overlap mechanics, though its claims omit variance and depend on undisclosed tuning that may fail with larger batches or mismatched gradients.
Abstract
This paper presents the design, implementation, and evaluation of the PyTorch distributed data parallel module. PyTorch is a widely-adopted scientific computing package used in deep learning research and applications. Recent advances in deep learning argue for the value of large datasets and large models, which necessitates the ability to scale out model training to more computational resources. Data parallelism has emerged as a popular solution for distributed training thanks to its straightforward principle and broad applicability. In general, the technique of distributed data parallelism replicates the model on every computational resource to generate gradients independently and then communicates those gradients at each iteration to keep model replicas consistent. Despite the conceptual simplicity of the technique, the subtle dependencies between computation and communication make it non-trivial to optimize the distributed training efficiency. As of v1.5, PyTorch natively provides several techniques to accelerate distributed data parallel, including bucketing gradients, overlapping computation with communication, and skipping gradient synchronization. Evaluations show that, when configured appropriately, the PyTorch distributed data parallel module attains near-linear scalability using 256 GPUs.