Method to purge a memory of objects, which are no longer accessible during program run

摘要

Programs in programming languages with dynamic storage space allocation require the memory to be purged from time to time of objects (e.g. datawords) which can no longer be accessed during the program run. Objects which can no longer be addressed (dead objects) are then made available again to the system. The method for purging the memory divides the address space into a first storage area (NS) and a second storage area (OS). In the first storage area, the new objects are stored in a first part-store (CA). During a first purging, the live objects of the first part-store (CA) are copied into a second part-store (FSS) of the first storage area. After the purging, the live objects saved in the second part-store (FSS) are recopied into a third part-store (PSS). During further purging processes, live objects are saved from the first part-store (CA) into the second part-store (FSS), live objects in the third part-store (PSS) are saved in the second storage area (OS). Objects newly saved in the second part-store (FSS) and live objects newly saved in the second storage area are alternately examined until no further live objects are found. The saved objects in the first storage area (NS) are then copied from the second part-store (FSS) into the third part-store (PSS). The second storage area (OS) is purged in accordance with the stop and copy method. IMAGE