Search papers, labs, and topics across Lattice.
This paper introduces MicroEvo, a novel framework that integrates large language models (LLMs) with Monte Carlo Tree Search (MCTS) to enhance the efficiency of microarchitecture design space exploration. By leveraging LLM-driven evolutionary operators and a Pareto-aware tree policy, MicroEvo significantly improves the quality of the Pareto front by up to 36.2% compared to traditional methods like NSGA-II, while also achieving a remarkable 10.6x increase in search efficiency. The approach incorporates active knowledge accumulation and state-aware directives to optimize search behavior dynamically, demonstrating strong scalability for complex industrial applications.
MicroEvo achieves a staggering 10.6x increase in search efficiency while improving Pareto-front quality by 36.2%, revolutionizing microarchitecture design exploration.
Microarchitecture design space exploration suffers from expansive search spaces and expensive PPA evaluation, leaving only a small simulation budget for design decision-making. Existing methods perform blind search without considering microarchitectural dependencies and fail to learn from the iterative search effectively, leading to wasted evaluations and weak Pareto convergence. In this paper, we propose MicroEvo, a knowledge-guided framework that couples off-the-shelf LLMs with Monte Carlo Tree Search (MCTS) for multi-objective microarchitecture optimization. MicroEvo combines LLM-driven evolutionary operators, a Pareto-aware tree policy that balances Pareto contribution and diversity, an active knowledge accumulation mechanism that extracts and reuses optimization insights, and state-aware directives that adapt the search behavior online. Experiments show that MicroEvo improves Pareto-front quality by up to 36.2% over NSGA-II and achieves 10.6x higher search efficiency, and also demonstrates strong scalability to a complex industrial-scale core. The code repository is available at: https://github.com/GEAR-SEU/MicroEvo-ICCAD-26.
NVIDIA