Unverified50% confidenceFactExact time
On UMA, the so-called 'on CPU' weights are actually sitting in the same unified RAM that the GPU also needs, and moving a tensor from MPS to CPU does not reduce overall memory pressure, only changes the access path and the state of the pages.
1
Sources
50%
Confidence
Long-term
Relevance
10/13/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedGPU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KB73% similarVerified苹果的统一内存架构（UMA）将CPU、GPU、神经网络引擎置于同一块内存池之上，消除了数据拷贝的开销73% similarUnverifiedWhen a model runs, its weights need to be loaded into RAM (system memory) or VRAM (video memory), and having large storage capacity does not mean the model can actually run.71% similarVerified统一内存架构使CPU、GPU与NPU共享同一块内存池，避免数据在处理单元间反复搬运的开销71% similarVerifiedThe GPU memory hierarchy from fastest to slowest is: registers, L1 cache/shared memory, L2 cache, global memory (VRAM).69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1015594API
curl https://kongchang.com/api/v1/knowledge/claims/1015594MCP
get_claim(id=1015594)