Unverified50% confidenceFactExact time
A100 GPU的HBM带宽约2TB/s，而片上SRAM带宽可达19TB/s，相差近10倍
1
Sources
50%
Confidence
Long-term
Relevance
7/20/2026
First Seen
Sources
Related Claims
UnverifiedGPU HBM带宽通常在2-3 TB/s量级，H100为3.35 TB/s，而PCIe 5.0连接的Host DRAM有效带宽仅约64 GB/s，两者相差近50倍83% similarPartially VerifiedA100 80GB采用HBM2e显存，带宽约2039GB/s；H100 SXM采用HBM3，带宽约3.35TB/s，FP64 Tensor Core算力约67 TFLOPS，是A100（约19.5 TFLOPS）的3倍以上81% similarUnverifiedA100的HBM2e带宽为2TB/s80% similarUnverified片上SRAM的访问带宽可达数百TB/s，而最先进的HBM3e仅能提供约8TB/s的带宽79% similarUnverifiedCerebras WSE-3片上内存高达44GB SRAM，带宽达到21 PB/s，比GPU的HBM内存快约1000倍77% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/566987API
curl https://kongchang.com/api/v1/knowledge/claims/566987MCP
get_claim(id=566987)