In a multicomputer parallel architecture, the shared database$/$dataset is "partitioned" across the local memories of the multicomputer processes. Which of the following are considerations for managing the performance of a multicomputer program? $($Pick Two$)$

Question $1$ options:

Each process in the multicomputer program relies mainly on its own local variables and its own local portion of the shared database,$$

Processes in a multicomputer architecture must be arranged in a Torus topology to minimize communications delays.

When a process in the multicomputer needs additional data it must use message passing. This message$-$passing must be relatively infrequent because of the long time delay involved in interprocessor communication.