Abstract: With the gap between the speed of processor and memory become wider and wider, memory access instructions especially frequently cause cache miss are the bottleneck of the performance. As the compiler does not know the exact cycles of memory access instruc

Key words: Scheduling, Cache profiling, MLP