Reliability Analysis and Assessment of Multi- Core System-on-Chip through Transaction Pro- filing and Machine Learning

Title	Reliability Analysis and Assessment of Multi- Core System-on-Chip through Transaction Pro- filing and Machine Learning
Summary	String representation "This project wi … f such systems." is too long.
Keywords
TimeFrame
References
Prerequisites
Author	Mahdi Fazeli
Supervisor	Mahdi Fazeli
Level	Master
Status	Open

In modern multi-core System-on-Chip (SoC) designs, ensuring reliability is critical as these systems are often deployed in environments that demand high performance and robustness against hardware faults or malicious attacks. Transaction profiling, which involves tracking the communication and data flow between cores, can provide valuable insights into the behavior of multi-core SoCs. By leveraging Machine Learning (ML) techniques, it is possible to analyze transaction data and identify patterns that may indicate reliability issues, potential security threats, or abnormal system behavior. The student will use a predeveloped toolchain to collect transaction data from prototype SoCs while running standard benchmarks such as Splash and Parsec. The collected transaction data, which represents the interactions between different cores and subsystems, will then be analyzed by ML models to detect any abnormal behaviors or system faults that compromise reliability. The key objectives of the project include: • Collecting and profiling transaction data from multi-core SoCs during benchmark execution with and withour faults being injected. • Training ML models to distinguish between normal and abnormal transactions, identifying signs of attacks or malfunctions. • Assessing the accuracy, efficiency, and scalability of the ML models in detecting reliability issues. • Providing insights into how transaction profiling can be used for real-time monitoring and predictive maintenance in SoC environments.