A disadvantage of the recursive structure in Figure 7.3.5 is that pipelining
within the PEs does not allow a direct increase in throughput
since the results of the current processing are required for the next processing step.
However, sequential data blocks of length N can be processed independently
of one another so that, in spite of the recursive structure, several data blocks
can be processed by interleaving (see section 4.1). The increased number of
intermediate registers in the recursive loop due to pipelining must correspond to the number of data blocks being processed by interleaving.