performance of the low-rank TT-SVD for large dense tensors on modern multicore CPUs