Function cublas_fp8_proj 

pub fn cublas_fp8_proj(
    gpu: &dyn GpuBackend,
    ptg_quant_k: KernelHandle,
    act_bf16: DevicePtr,
    act_fp8_scratch: DevicePtr,
    act_scale_scratch: DevicePtr,
    fp8w: &Fp8Weight,
    out: DevicePtr,
    m: u32,
    n: u32,
    k: u32,
    stream: u64,
) -> Result<()>

Route a projection through native-FP8 cuBLASLt block-scaled matmul: quantize the activation to FP8 + per-[token,128-of-K] VEC128 scales (the existing per_token_group_quant_fp8 kernel), feed the FP8 weight + its per-128×128 block scales directly (zero dequant, zero extra weight memory). Both operands 128-block-scaled (cuBLASLt requires it). ~1.8× the bf16 path (152 vs 85 TF).

act_fp8_scratch/act_scale_scratch must hold the padded extents (the buffers.fp8_act/fp8_act_scale arena buffers, sized for max_batch_tokens).