ConceptAI & ML
TensorFloat-32
TF32 is a math mode for NVIDIA Ampere architecture GPUs that uses FP32 inputs and outputs with an internal 19-bit precision representation to accelerate matrix multiplications.
Know first
Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.
Lecturers
1Connections
3The lectures that define it or assume it, and the concepts next to it.