Expired90% confidenceFactTime unknown
On 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.
7
Sources
90%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
Related Claims
UnverifiedNVIDIA's H100 GPU natively supports the FP8 format, with throughput approximately 2× that of FP16.84% similarUnverifiedConverting from FP32 to FP16 can achieve a 2–4x throughput improvement on the Tensor Cores of modern GPUs.84% similarUnverifiedThe NVIDIA Blackwell architecture (B200/GB200) supports FP4 computation, theoretically doubling Tensor Core throughput once again.80% similarUnverifiedNVIDIA designed a dedicated Tensor Core instruction mma.kind::mxf4 for NVFP4 in the Blackwell architecture (SM 100/103, such as the B100 and B200 series), and adopted an FP8 scaling factor mechanism with one scale per group of 16 elements.80% similarVerifiedThe NVIDIA A100 has 6,912 CUDA cores and 432 Tensor Cores, with a theoretical peak compute of 312 TFLOPS (BF16).79% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/61143API
curl https://kongchang.com/api/v1/knowledge/claims/61143MCP
get_claim(id=61143)