Expired70% confidenceBenchmarkExact time
NVIDIA's H100 GPU natively supports the FP8 format, with throughput approximately 2× that of FP16.
2
Sources
70%
Confidence
Medium-term (~90 days)
Relevance
7/7/2026
First Seen
Valid until: 10/5/2026(expired)
Sources
NVIDIA AI-Q蓝图×OCI：企业级AI智能体生产部署实践指南
rss6/26/2026
Related Claims
VerifiedOn NVIDIA GPUs equipped with Tensor Cores (such as A100 and H100), the throughput of FP16/BF16 matrix operations can reach 2–8 times that of FP32.84% similarUnverified现代GPU如NVIDIA H100、H800原生支持FP8计算，理论上可实现约2倍吞吐量提升79% similarPartially VerifiedNVIDIA's MIG technology enables hardware-level resource isolation on high-end GPUs such as the A100/H100.78% similarUnverifiedNVIDIA's iteration cycle from A100 to H100 and then to B100 is approximately two years, with each generation delivering roughly a 3–4× improvement in FP8 mixed-precision compute.77% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/185584API
curl https://kongchang.com/api/v1/knowledge/claims/185584MCP
get_claim(id=185584)