When a developer deploys an AI model using an NVIDIA Inference Microservice (NIM) on HPE Private Cloud AI, the NIM automatically detects the underlying NVIDIA GPU hardware.
The developer sees the following status message upon deployment:
```
NIM Status:
- Model: mistral-7b-instruct-v0.2
- Status: Running
- Engine Profile: Selected 'TensorRT-LLM FP8' for optimal latency on detected Hopper-class GPU.
```
What does the selection of the 'TensorRT-LLM FP8' engine profile indicate?