Expired80% confidenceFactTime unknown
The NVIDIA Hopper architecture (H100/H200) introduced FP8 Tensor Cores for the first time and added the Transformer Engine to automatically manage precision conversion.
4
Sources
80%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
Related Claims
UnverifiedNative FP8 hardware acceleration was first implemented in the fourth-generation Tensor Cores of the NVIDIA Hopper architecture (H100, 2022), with a peak throughput of 3958 TFLOPS.80% similarVerifiedNVIDIA's H100 features a Transformer Engine specifically designed for AI workloads, delivering nearly 4,000 TFLOPS of compute at FP8 precision, approximately 3 times the improvement over the A100.78% similarUnverifiedThe Hopper architecture of the H100 introduced a dedicated Transformer Engine supporting FP8 precision computation.77% similarUnverifiedNVIDIA B200采用Blackwell架构，于2024年发布，是H100的继任者，引入第五代NVLink、支持FP4精度的新一代Tensor Core以及Transformer Engine 2.076% similarUnverifiedNVIDIA Ampere 架构引入了异步内存拷贝和稀疏矩阵加速单元，Hopper 架构引入了 Transformer Engine 专用硬件75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/61135API
curl https://kongchang.com/api/v1/knowledge/claims/61135MCP
get_claim(id=61135)