Problem of the optimization of memory hierarchies in automated control systems

Conclusion

Thus the algorithm described allows us to construct a set of optimal hierarchical memory structures for an actual information system and to investigate their cost-time parameters.

When practical use is made of the method of optimizing hierarchical memory structures, it is necessary to take into account certain features of real information-storage systems. These include, in particular, the following factors:

The maximum number of homogeneous blocks of certain memory types may be restricted by the possibilities of the corresponding communication channels to the processor.

As the number of modules of given types of memory is increased by some step, an additional control device is needed (when the number of blocks is decreased, the converse applies) so that the overall cost will vary not in direct proportion to the number of modules.

For a serial-access memory (magnetic-tape-store type), the cost-time results of the transfer of information with highest activity into the major types of memory requires a more detailed analysis based on taking into account the variation of the location structure of actual information elements.

These and certain other factors may be taken into account within the framework of the proposed method by making the algorithm correspondingly more complicated without changing its essence.