PURPOSE: To omit the unnecessary data identifiers by detecting the data that should be transferred between the processors and allocating the data identifiers to these extracted data only.
CONSTITUTION: The data identifiers are generated for identification of data which are transferred among plural processors. Then a data group including the different processors for use and production of data is extracted based on the data using range information showing a range of data used by each processor and the data generating range information showing a range of data produced by each processor. The data identifiers are allocated to only the data included in the extracted data group. Thus it is possible to minimize a possible range of data identifiers to be allocated to the data transferred among parallel processors. As a result, the data can be transferred among plural processors asynchronously with each other and with the use of a small number of data identifiers.
OKOCHI TOSHIO
KONNO CHISATO
HITACHI VLSI ENG