Puzzle: Scheduling Multiple Deep Learning Models on Mobile Device with Heterogeneous Processors
By: Duseok Kang, Yunseong Lee, Junghoon Kim
Potential Business Impact:
Lets phones run many AI tasks faster.
As deep learning models are increasingly deployed on mobile devices, modern mobile devices incorporate deep learning-specific accelerators to handle the growing computational demands, thus increasing their hardware heterogeneity. However, existing works on scheduling deep learning workloads across these processors have significant limitations: most studies focus on single-model scenarios rather than realistic multi-model scenarios, overlook performance variations from different hardware/software configurations, and struggle with accurate execution time estimation. To address these challenges, we propose a novel genetic algorithm-based methodology for scheduling multiple deep learning networks on heterogeneous processors by partitioning the networks into multiple subgraphs. Our approach incorporates three different types of chromosomes for partition/mapping/priority exploration, and leverages device-in-the-loop profiling and evaluation for accurate execution time estimation. Based on this methodology, our system, Puzzle, demonstrates superior performance in extensive evaluations with randomly generated scenarios involving nine state-of-the-art networks. The results demonstrate Puzzle can support 3.7 and 2.2 times higher request frequency on average compared to the two heuristic baselines, NPU Only and Best Mapping, respectively, while satisfying the equivalent level of real-time requirements.
Similar Papers
Scheduling Techniques of AI Models on Modern Heterogeneous Edge GPU -- A Critical Review
Distributed, Parallel, and Cluster Computing
Makes smart gadgets run AI faster and better.
Resource Heterogeneity-Aware and Utilization-Enhanced Scheduling for Deep Learning Clusters
Distributed, Parallel, and Cluster Computing
Makes computer learning faster and better.
Accelerating Mobile Inference through Fine-Grained CPU-GPU Co-Execution
Machine Learning (CS)
Lets phones run smart programs much faster.