Unverified50% confidenceFactExact time
PagedAttention 由 vLLM 团队于 2023 年提出，借鉴操作系统虚拟内存分页机制，将 KV Cache 内存浪费率从超过 60% 降低至接近 4%
1
Sources
50%
Confidence
Long-term
Relevance
7/21/2026
First Seen
Sources
Related Claims
UnverifiedvLLM's PagedAttention draws on the operating system concept of virtual memory paging to manage KV caches, reducing GPU memory fragmentation overhead from approximately 30% to near 0.84% similarUnverifiedvLLM compresses the KV Cache memory fragmentation rate from above 60% to near zero through PagedAttention technology, and can increase inference throughput by several times in high-concurrency scenarios.80% similarUnverifiedvLLM的PagedAttention将KV Cache碎片化率从传统方案的60-80%降至不足4%，吞吐量提升可达24倍73% similarUnverifiedvLLM的PagedAttention将KV Cache划分为固定大小物理块，通过块表实现地址映射，将碎片化率从60-80%降至不足4%，吞吐量提升可达24倍72% similarUnverifiedCompared to traditional methods, PagedAttention enables vLLM to increase batch size by 2–4 times and achieve close to 100% memory utilization.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/575978API
curl https://kongchang.com/api/v1/knowledge/claims/575978MCP
get_claim(id=575978)