SparDL: Distributed Deep Learning Training with Efficient Sparse Communication
This addresses communication bottlenecks in distributed training for ML practitioners, though it appears incremental as it builds on existing sparse communication techniques.
The paper tackles the Sparse Gradient Accumulation (SGA) dilemma in distributed deep learning with top-k sparsification, proposing SparDL, which achieves up to 1.98× speedup over state-of-the-art methods.
Top-k sparsification has recently been widely used to reduce the communication volume in distributed deep learning. However, due to the Sparse Gradient Accumulation (SGA) dilemma, the performance of top-k sparsification still has limitations. Recently, a few methods have been put forward to handle the SGA dilemma. Regrettably, even the state-of-the-art method suffers from several drawbacks, e.g., it relies on an inefficient communication algorithm and requires extra transmission steps. Motivated by the limitations of existing methods, we propose a novel efficient sparse communication framework, called SparDL. Specifically, SparDL uses the Spar-Reduce-Scatter algorithm, which is based on an efficient Reduce-Scatter model, to handle the SGA dilemma without additional communication operations. Besides, to further reduce the latency cost and improve the efficiency of SparDL, we propose the Spar-All-Gather algorithm. Moreover, we propose the global residual collection algorithm to ensure fast convergence of model training. Finally, extensive experiments are conducted to validate the superiority of SparDL.