Unverified50% confidenceFactExact time
GPU上下文切换需保存完整GPU状态（包括寄存器文件、共享内存和L1缓存），数据量可达数百MB，而CPU上下文切换仅需保存几KB
1
Sources
50%
Confidence
Long-term
Relevance
7/16/2026
First Seen
Sources
Related Claims
UnverifiedThe GPU memory hierarchy from fastest to slowest is: registers, L1 cache/shared memory, L2 cache, global memory (VRAM).82% similarUnverified当专用GPU显存不够用时，Windows系统会自动将部分数据转移到从系统RAM划拨的共享GPU内存中79% similarUnverifiedThe GPU multi-level memory hierarchy from highest to lowest is: register file (latency ~1 cycle), shared memory/L1 cache (~32–228 KB, latency ~20–40 cycles), L2 cache (H100 ~50 MB, latency ~200 cycles), and HBM global memory (latency ~500–800 cycles).78% similarUnverifiedGPU的SM能同时驻留的Warp数量决定了GPU隐藏内存访问延迟的能力，制约Occupancy的核心资源是寄存器文件与共享内存77% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/527693API
curl https://kongchang.com/api/v1/knowledge/claims/527693MCP
get_claim(id=527693)