Memory mapping method in a data processing system

Abstract

In a data processing system wherein a memory is comprised of an unknown plurality of memory blocks of a basic capacity, arranged in an unknown plurality of modules which have an unknown capacity multiple of the basic capacity, a method addresses the memory location which involves the selection of the module containing such locations by use of a directory having a plurality of addressable locations. The initial loading of the directory with codes corresponding to the real constituent of the memory is performed by writing into the directory locations a binary code assigning the related memory blocks to a hypothetical first module, writing test codes into memory locations each belonging to a different block, reading out the contents of the same memory locations and comparing such contents with the test codes to determine if a first module is present and if such memory locations and the pertaining blocks belong to the first module, writing into the directory locations whose related memory block do not belong to the first module a binary code assigning the related memory blocks to a hypothetical second module, writing test codes into memory locations each belonging to a different block assigned to the second module, reading out the contents of the same memory locations and comparing such contents with the test codes to determine if a second module is present and if such memory locations and related memory blocks belong to the second module, and repeating the above operations for further subsequent modules.