94.574 Elements of Computer Systems

SCE 574

226

Some Observations

äThe overhead to keep track of small holes will be substantially larger than the hole itself. äThe best-fit is slower than the first-fit because it must search the entire list. äThe best-fit tends to fill up memory with tiny useless holes.

äThe first-fit generates larger holes on average

äSimulation have shown that the first-fit and the best-fit give better results, with the first-fit being faster. äTo speed up these algorithms, we can maintain separate lists for processes and holes äTo make the best-fit faster, we can maintain the hole list by size.