JExplore: Design Space Exploration Tool for Nvidia Jetson Boards
Basar Kutukcu, Sinan Xie, Sabur Baidya, Sujit Dey
TL;DR
JExplore tackles the challenge of navigating the large configurability space of Nvidia Jetson boards for AI workloads by providing a reusable host–client framework that integrates with arbitrary search algorithms. It encapsulates hardware configuration management, measurement, and result logging, and enables batch sampling across multiple Jetson devices via ZMQ over SSH. The authors demonstrate the tool with two real-world AI workloads (Llama2-7B and LLaVA-1.5-7B), revealing a clear power–latency trade-off and EMC-frequency effects that shape the design space frontier. The work delivers an open-source benchmarking ground and API surface to accelerate the development and comparison of search strategies in embedded AI hardware contexts.
Abstract
Nvidia Jetson boards are powerful systems for executing artificial intelligence workloads in edge and mobile environments due to their effective GPU hardware and widely supported software stack. In addition to these benefits, Nvidia Jetson boards provide large configurability by giving the user the choice to modify many hardware parameters. This large space of configurability creates the need of searching the optimal configurations based on the user's requirements. In this work, we propose JExplore, a multi-board software and hardware design space exploration tool. JExplore can be integrated with any search tool, hence creating a common benchmarking ground for the search algorithms. Moreover, it accelerates the exploration of user application and Nvidia Jetson configurations for researchers and engineers by encapsulating host-client communication, configuration management, and metric measurement.