Unverified50% confidenceFactExact time

自回归模型随序列长度增加，KV Cache占用线性增长，显存带宽成为制约速度的硬件瓶颈

1

Sources

50%

Confidence

Long-term

Relevance

7/13/2026

First Seen

Sources

扩散语言模型崛起：DiffusionGemma实测速度碾压自回归模型

redditr/GoogleGeminiAI7/11/2026

Related Claims

Unverified卸载方案下提示处理速度和内存带宽往往成为限制因素，PCIe 带宽瓶颈会导致推理吞吐量急剧恶化75% similar Unverified超长上下文窗口的KV缓存显存占用随上下文长度线性增长，导致推理延迟和成本显著上升74% similar Unverified长文档处理中，KV Cache机制导致显存占用随上下文长度近似平方级增长69% similar Verified移动端模型推理的速度瓶颈往往不在算力而在内存带宽69% similar Unverified在高缓存命中率场景下盲目增加GPU可能收效甚微，应先识别带宽瓶颈再优化69% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/497403

API

curl https://kongchang.com/api/v1/knowledge/claims/497403

MCP

get_claim(id=497403)