Feature Request: support FP8 data type in llama.cpp

[wenlujon]

Prerequisites

• I am running the latest code. Mention the version if possible as well.
• I carefully followed the README.md.
• I searched using keywords relevant to my issue to make sure that I am creating a new issue that is not already open (or closed).
• I reviewed the Discussions, and have a new and useful enhancement to share.

Feature Description

Currently, the FP8 data type is the default data type for models like DeepSeek, which is supposed to have better accuracy than INT8 while they have the same size. vLLM already supports fp8 data type.

File the issue for request the support of fp8 in llama.cpp.

Motivation

with same level model size, FP8 is supposed to have better accuracy than int8.

Possible Implementation

No response