| 摘要 |
Efficient memory area allocation is performed in an error free manner when a plurality of threads access memory areas in parallel. A thread list of thread information containing entry time information that is recorded on a per thread basis with a thread being as a data processing unit is stored as thread management information, and release queue containing release request time that is recorded on an area unit basis concerning a release-requested but not yet released memory area is stored as memory area management information. The release request time set in each queue component contained in the release queue is compared with the oldest entry time of each queue component in the thread list during a memory area allocation process. The memory allocation process is performed on only the memory area corresponding to the queue component with the release request time set prior to the oldest entry time. In this arrangement, a memory area that is not accessed is reliably selected for allocation.
|
