CACHEFORGE: LLM-Guided End-to-End Generative Cache Replacement Policy for Performance and Hardware Efficiency
CACHEFORGE is the first framework to evolve cache-replacement policies end-to-end by embedding a large language model inside a governed hardware-aware loop and enforces feasibility through reward shaping, structural checks, dynamic mutation, temperature scheduling, and cross-policy crossover.
Kaushal Mhapsekar, Bita Aslrousta, Brijesh Kumar Bhayana et al.
· 0 citations