This document is relevant for: Inf2, Trn1, Trn1n

nki#

Framework#

Simulate a nki kernel on CPU using a built-in simulator in Neuron Compiler.

`benchmark`	Benchmark a NKI kernel on a NeuronDevice by using `nki.benchmark` as a decorator.
`baremetal`	Compile and run a NKI kernel on NeuronDevice without involving ML frameworks such as PyTorch and JAX.

NKI kernels are represeted as XLA CustomCall instructions in HLO.

A tensor object represents a multidimensional, homogeneous array of fixed-size items

This document is relevant for: Inf2, Trn1, Trn1n