Boost Fine-Tuning Performance of LLM: Optimal Architecture w/ PEFT LoRA Adapter-Tuning on Your GPU