Echo: Error-Compensated 4-Bit Tensor-Parallel Communication for LLM Training and Inference

본문 미리보기

arXiv:2610.14277v1 Announce Type: new Abstract: Tensor parallelism ends every layer with an all-reduce that sits between two dependent matrix multiplications, so its cost cannot be hidden and grows as compute outpaces interconnect bandwidth. Sending 4-bit instead of 16-bit values would cut this cost by nearly 4x, but the all-reduce must add the values it carries, and rounding the sum injects an error at every call. Model quality depends on how these errors accumulate over layers, yet existing m