thinQmaster Virtual testing partner Adaptive practice exams that find your weak spots and drill them. Every answer explained. From $69
Enroll

thinQmaster

Find your exam

Five questions free, before you decide.

Step 1 · choose a vendor

In one of our camps? Your practice exams come with the tuition — they are on your study page already.

AI & Machine Learning · Advanced

FP8

An 8-bit floating-point format supported by NVIDIA Hopper and later Tensor Cores for faster training and inference.

Definition

FP8 comes in E4M3 (more precision) and E5M2 (more range) variants. NVIDIA's Transformer Engine manages per-tensor scaling so models can train and infer in FP8 with minimal accuracy loss, roughly doubling throughput over 16-bit.

Exam tip

FP8 support arrived with Hopper (H100); FP4 arrived with Blackwell.

Related terms