Evaluating the Applicability of LLM Inference Optimizations on Apple Silicon