Revisiting Distributed Synchronous SGD
This work addresses the problem of efficient distributed training for deep learning practitioners, offering a practical improvement over existing approaches.
The paper tackles the trade-off between asynchronous and synchronous distributed SGD by proposing synchronous optimization with backup workers, which converges faster and achieves better test accuracies than conventional methods.
Distributed training of deep learning models on large-scale training data is typically conducted with asynchronous stochastic optimization to maximize the rate of updates, at the cost of additional noise introduced from asynchrony. In contrast, the synchronous approach is often thought to be impractical due to idle time wasted on waiting for straggling workers. We revisit these conventional beliefs in this paper, and examine the weaknesses of both approaches. We demonstrate that a third approach, synchronous optimization with backup workers, can avoid asynchronous noise while mitigating for the worst stragglers. Our approach is empirically validated and shown to converge faster and to better test accuracies.