[feat] FP8 Matmul Training Kernel

### 🚀 The feature, motivation and pitch

FP8 Training has been a great weapon on H100 and provides huge memory and speed benefits, and has shown to be effective (with limited or no performance drop) in many cases and reports. The kernel will be a fused fp8 matmul kernel with fusion of three parts:

input quantization (dynamic, static) + fp8 matmul (weight dynamic, static quantization inside) + output dequantization

### Alternatives

_No response_

### Additional context

_No response_

The issue names no files, tests, or entry points. Search the repository for existing matmul and quantization kernels, then clarify the supported dynamic and static FP8 paths and their expected interfaces. Done means a fused kernel covers input quantization, FP8 matmul with weight quantization, and output dequantization, with validation of correctness and performance.

Written by the indexing model from the issue text.