Performance Isolation for Inference Processes in Edge GPU Systems
For developers of safety-critical systems using deep learning on shared GPUs, this study provides a comparative analysis of isolation mechanisms to ensure predictable inference times.
This work evaluates GPU isolation mechanisms (MPS, MIG, Green Contexts) for predictable inference in safety-critical applications, finding MIG provides high isolation while Green Contexts offer fine-grained SM allocation with low overhead but no memory isolation.
This work analyzes the main isolation mechanisms available in modern NVIDIA GPUs: MPS, MIG, and the recent Green Contexts, to ensure predictable inference time in safety-critical applications using deep learning models. The experimental methodology includes performance tests, evaluation of partitioning impact, and analysis of temporal isolation between processes, considering both the NVIDIA A100 and Jetson Orin platforms. It is observed that MIG provides a high level of isolation. At the same time, Green Contexts represent a promising alternative for edge devices by enabling fine-grained SM allocation with low overhead, albeit without memory isolation. The study also identifies current limitations and outlines potential research directions to improve temporal predictability in shared GPUs.