Presentation
A Deep Reinforcement Learning based Online Scheduling Policy for Deep Neural Network Multi-Tenant Multi-Accelerator Systems
DescriptionDeep Learning, particularly Deep Neural Networks (DNNs), has emerged as a powerful tool for addressing intricate real-world challenges. Nonetheless, the deployment of DNNs presents its own set of obstacles, chiefly stemming from substantial hardware demands. In response to this challenge, Domain-Specific Accelerators (DSAs) have gained prominence as a means of executing DNNs, especially within cloud service providers offering DNN execution as a service. For service providers, managing multi-tenancy and ensuring high quality service delivery, particularly in meeting stringent execution time constraints, assumes paramount importance, all while endeavoring to maintain cost-effectiveness. In this context, the utilization of heterogeneous multi-accelerator systems becomes increasingly relevant. This paper presents RELMAS, a low-overhead deep reinforcement learning algorithm designed for the real-time scheduling of DNNs in multi-tenant environments, taking into account the dataflow heterogeneity of accelerators and memory bandwidths contentions. By doing so, service providers can employ the most efficient scheduling policy for user requests, optimizing Service-Level-Agreement (SLA) satisfaction rates and enhancing hardware utilization. The application of RELMAS to a heterogeneous multi-accelerator system composed of various instances of Simba and Eyeriss sub-accelerators resulted in up to a 173% improvement in SLA satisfaction rate compared to state-of-the-art scheduling techniques across different workload scenarios, with less than a 1.5% energy overhead.
Event Type
Research Manuscript
TimeThursday, June 2711:30am - 11:45am PDT
Location3010, 3rd Floor
AI
Design
AI/ML Architecture Design