NuPS: A Parameter Server for Machine Learning with Non-Uniform Parameter Access
This addresses performance bottlenecks in distributed training for machine learning practitioners, offering a significant improvement over existing systems.
The paper tackled the inefficiency of existing parameter servers in distributed machine learning tasks with non-uniform parameter access, such as skew and sampling, by introducing NuPS, a novel architecture that integrates multiple management techniques and supports sampling primitives. The result was that NuPS outperformed existing parameter servers by up to one order of magnitude and achieved up to linear scalability across tasks.
Parameter servers (PSs) facilitate the implementation of distributed training for large machine learning tasks. In this paper, we argue that existing PSs are inefficient for tasks that exhibit non-uniform parameter access; their performance may even fall behind that of single node baselines. We identify two major sources of such non-uniform access: skew and sampling. Existing PSs are ill-suited for managing skew because they uniformly apply the same parameter management technique to all parameters. They are inefficient for sampling because the PS is oblivious to the associated randomized accesses and cannot exploit locality. To overcome these performance limitations, we introduce NuPS, a novel PS architecture that (i) integrates multiple management techniques and employs a suitable technique for each parameter and (ii) supports sampling directly via suitable sampling primitives and sampling schemes that allow for a controlled quality--efficiency trade-off. In our experimental study, NuPS outperformed existing PSs by up to one order of magnitude and provided up to linear scalability across multiple machine learning tasks.