Function deinterleave_qg_split 

pub fn deinterleave_qg_split(
    gpu: &dyn GpuBackend,
    kernel: KernelHandle,
    data: DevicePtr,
    q_out: DevicePtr,
    num_tokens: u32,
    num_heads: u32,
    head_dim: u32,
    stride: u32,
    stream: u64,
) -> Result<()>

2026-09-25: deinterleave_qg with Q written to q_out (contiguous [num_tokens, num_heads * head_dim]) instead of in place. The gate is written back into data at offset num_heads * head_dim of each row.