Verified75% confidenceFactExact time
NVIDIA H20 的 BF16 峰值算力约为 296 TFLOPS，相比 H100 的 989 TFLOPS 大幅削减，但保留了 141GB HBM3 显存与 4.0 TB/s 显存带宽
5
Sources
75%
Confidence
Long-term
Relevance
7/16/2026
First Seen
Sources
Related Claims
VerifiedNVIDIA H100的HBM显存带宽约为3.35 TB/s，在自回归生成任务中利用率往往不足10%83% similarVerifiedNVIDIA A100单卡拥有80GB HBM2e显存和312 TFLOPS的FP16算力80% 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76% similarVerifiedNVIDIA H100单卡FP64双精度浮点性能达到约34 TFLOPS76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/533867API
curl https://kongchang.com/api/v1/knowledge/claims/533867MCP
get_claim(id=533867)