Re: LRU and full table scans

Roque Bonilla wrote:
> First Hello, Second...Yes, I'm doing some studies on the buffer manager and
> also studying and implementing different policys, one of them is the LRU-K.
>
> By the time I don't have any results, I'm just preparing to run the TPC-H
> benchmark to test all the policys that I have implemented.
> I have implemented some policys for the buffer manager, some better and some
> worst than LRU (FIFO, LFU, LRD, FBR, LRU-K, LRFU, 4 CLOCK policys (or second
> chance), CORRELATED REFERENCES, and different combinations of them like
> (LFU+CLOCK+AGING(by division or by substraction)+CORRELATED REFERENCES)),
> there are seven principal policys and seven "add-on's" that could be applyed
> to them, resulting in 25 combinations of policys.

We do have someone working on LRU-K but we haven't seen any test results
from him yet. Please let us know what you find.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026