Basic Model

Figure 3: Communication and computation in a star interconnection.

The process of data distribution and computations for a star of PEs with network processors, without returning the results is depicted in Fig.3. It can be observed that the time of computing on $P_i$ equals to the sum of the time required for sending the load to $P_{i+1}$ and computing on $P_{i+1}$. Thus, distribution of the load can be found from formulae (1) where the first line is replaced by:

$$\alpha_iA_i = S_{i+1}+\alpha_{i+1}(C_{i+1}+A_{i+1}), {\rm\hspace{3mm}} i=1,\dots,m-1$$ (5)

This set of equations can be solved analogously to (1).