Unverified80% confidenceFactTime unknown
Blackwell架构支持FP4精度计算，FP4是目前商用AI芯片支持的最低精度格式
1
Sources
80%
Confidence
Long-term
Relevance
5/31/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe Blackwell architecture introduces fifth-generation Tensor Cores with native support for FP4×FP4 matrix multiply-accumulate operations, using FP32 or FP16 precision for the accumulator.82% similarUnverifiedBlackwell的FP4格式仅有4位（1位符号、2位指数、1位尾数），采用每128个FP4元素共享一个FP8精度缩放因子的块浮点表示75% similarUnverifiedFP8 is a low-precision computing format natively supported by the Blackwell architecture, including two variants—E4M3 and E5M2—and the Tensor Core's FP8 compute throughput is twice that of FP16.73% similarUnverifiedBlackwell is NVIDIA's latest-generation data center architecture (B100/B200 series), featuring a dual-chip package design (MCM), and a single GPU can deliver up to 20 petaFLOPS of FP4 compute performance.73% similarUnverifiedBlackwell introduces FP4 for inference, and combined with microscaling technology, can compress weight storage to 1/4 of FP16.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/4497API
curl https://kongchang.com/api/v1/knowledge/claims/4497MCP
get_claim(id=4497)