EasyScale: Elastic Training with Consistent Accuracy and Improved Utilization on GPUs
Citations Over TimeTop 10% of 2023 papers
Abstract
Distributed synchronized GPU training is commonly used for deep learning. The resource constraint of using a fixed number of GPUs makes large-scale training jobs suffer from long queuing time for resource allocation, and lowers the cluster utilization. Adapting to resource elasticity can alleviate this but often introduces inconsistent model accuracy, due to lacking of capability to decouple model training procedure from resource allocation. We propose EasyScale, an elastic training system that achieves consistent model accuracy under resource elasticity for both homogeneous and heterogeneous GPUs. EasyScale preserves the data-parallel training behaviors strictly, traces the consistency-relevant factors carefully, utilizes the deep learning characteristics for EasyScaleThread abstraction and fast context-switching. To utilize heterogeneous cluster, EasyScale dynamically assigns workers based on the intra-/inter-job schedulers, minimizing load imbalance and maximizing aggregated job throughput. Deployed in an online serving cluster, EasyScale powers the training jobs to utilize idle GPUs opportunistically, improving overall cluster utilization by 62.1%.
Related Papers
- Comparison of farmers workload by manual and mechanical tasks on family farms.(2013)
- The idle thoughts of an idle fellow, a book for an idle holiday(2010)
- → Workload Transitions in Driving(2008)
- Design and Implementation of Teacher Workload Management System in University(2010)
- → ‘Unite Idle Men with Idle Land’:(2016)