Search papers, labs, and topics across Lattice.
Multi-core hardware interference analysis is framed as an active state-space exploration problem to overcome the blind spots of manual and pseudo-random micro-architectural testing in safety-critical systems. This matters because shared resource contention undermines deterministic execution time guarantees required by avionics and embedded certification standards. Evaluated in a simulated multi-core environment, intrinsic curiosity-driven exploration agents achieve substantially broader and more uniform behavioral coverage of complex interference states under strict evaluation budgets compared to traditional pseudo-random workload generators.
Standard random fuzzing misses critical multi-core hardware bottlenecks, but treating micro-architectural stress-testing as an intrinsic curiosity problem reliably surfaces elusive, worst-case contention edge cases under fixed experimental budgets.
The transition from single-core to multi-core architectures in safety-critical embedded systems introduces significant challenges due to inter-core interference caused by contention for shared hardware resources. Such interference affects execution times and complicates the verification of strict temporal requirements, particularly in domains such as avionics where standards require comprehensive identification of interference sources. Existing interference analysis approaches, whether manual or model-based, struggle to capture the full range of behaviors arising from the complex interactions among micro-architectural components. In this paper, we frame multi-core interference analysis as the exploration of a complex system behavior space. We propose the use of curiosity-driven exploration algorithms from artificial intelligence to systematically and efficiently cover the space of possible interference behaviors. Using a simulator-based environment, we show that the proposed approach achieves broader and more uniform behavioral coverage within a limited experimental budget compared to traditional pseudo-random program generation methods.
