Function gemm_with 

pub fn gemm_with<T: GemmScalar>(
    ws: &mut Workspace,
    alpha: T,
    a: MatRef<'_, T>,
    b: MatRef<'_, T>,
    beta: T,
    c: MatMut<'_, T>,
    par: Parallelism,
)

Like gemm but reuses a caller-owned Workspace: zero heap allocation once the workspace has grown to fit the 1st sufficiently large call

Panics

Same conditions as gemm