斯坦福大学 CS336 Lecture 05 GPU Principles and Distributed Training Fundamentals
1. Goals①. Having a thorough understanding of the working principle and operation mechanism of GPU ②. Knowing how to make fast algorithms like FlashAttention2. GPU Architecture2.1 Why ParallelismDennard Scaling:随着晶体管变得越来越小ÿ…
2026/9/18 13:22:21