|
|
|
|
|
|
by squeed
4312 days ago
|
|
|
2Q is clever, but certainly not the only improvement over LRU for page caching. Of interest, also, is LIRS, which uses the timing between the last two accesses to decide on a caching strategy. There is a good comparison of page caching algorithms here: http://people.cs.vt.edu/~butta/docs/sigmetrics05_kernelPrefe... . However, the conclusion seems to be that for non-synthetic loads, the difference is fairly minimal. |
|
|