Unverified50% confidenceFactExact time
冯·诺依曼瓶颈指处理器和存储之间的数据传输带宽限制
1
Sources
50%
Confidence
Long-term
Relevance
8/6/2026
First Seen
Sources
Related Claims
Unverified冯·诺依曼架构的核心特征是计算单元与存储单元分离，数据需在二者间搬运，造成能耗主要来源于数据传输而非计算本身75% similarUnverified主机卸载用带宽换显存空间，代价是数据传输延迟；激活重计算用算力换显存，二者可互补配合72% similarVerifiedThe speed bottleneck in large model inference is memory bandwidth rather than insufficient computing power, as generating each token requires loading hundreds of GB of model weights from memory.67% similarUnverifiedAutoregressive text generation is a memory-bandwidth-bound task, where memory bandwidth—not peak compute—is the true performance bottleneck.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/694295API
curl https://kongchang.com/api/v1/knowledge/claims/694295MCP
get_claim(id=694295)