Abstract: When massive spatial data are processed with computer, the data exchange between external memory and internal memory is one of the key factors affecting the efficiency of system. This paper firstly analyzes the drawbacks of system-level page replacement algorithm, then proposes the concept of application-level page replacement, and also describes in detail its implementation in Windows operating system. This algorithm is validated in face recognition system, the results show that it performs better than the original algorithm in efficiency greatly.

Key words: page replacement algorithm, memory pool, address windowing extension