Function bgemm_strided_into 

pub fn bgemm_strided_into<T>(
    c: &mut StridedViewMut<'_, T>,
    a: &StridedView<'_, T>,
    b: &StridedView<'_, T>,
    _n_batch: usize,
    n_lo: usize,
    n_ro: usize,
    n_sum: usize,
    alpha: T,
    beta: T,
    conj_a: bool,
    conj_b: bool,
) -> Result<()>
where
    T: ComplexField + Copy + ElementOpApply + Send + Sync + Mul<Output = T> + Add<Output = T> + Zero + One + PartialEq,

Batched strided GEMM using faer: C = alpha * A * B + beta * C

Same interface as bgemm_naive::bgemm_strided_into. Uses faer’s optimized matmul for all cases. When dimension groups have non-contiguous strides, copies operands to contiguous column-major buffers first using strided_kernel::copy_into.