Louis Aeilot's Blog
blog.aeilot.top
GPUs Modern AI training is structured as a hierarchy: individual GPUs (or TPUs) sit inside servers, servers are grouped into pods, pods into racks, and racks form a cluster. Instead of training on a s…
Louis Aeilot's Blog 的其他文章
- How Close Is FlashAttention to the Limit? Understanding Attention Through Data Movement
- Beyond FLOPs: How COSMA Builds Parallel Matrix Multiplication from Communication Bounds
- The Red-Blue Pebble Game: Why Faster Processors Still Have to Move Data
- Git Needs a Trash Can
- RoPE: Properties, Patterns, and Long-Context Behavior
讨论
登录后参与讨论
还没有评论，来说第一句吧。