Large Language Model (LLM) inference faces a fundamental challenge: the same hardware that excels at processing input prompts ...