Re: [HACKERS] Some notes on optimizer cost estimates

Then <tgl(at)sss(dot)pgh(dot)pa(dot)us> spoke up and said:
> As best I can measure on my hardware, the cost of a nonsequential
> disk read should be estimated at 4 to 5 times the cost of a sequential
> one --- I'm getting numbers like 2.2 msec per disk page for sequential
> scans, and as much as 11 msec per page for index scans. I don't
> know, however, if this ratio is similar enough on other platforms
> to be useful for cost estimating. We could make it a parameter like
> we do for CPU_PAGE_WEIGHT ... but you know and I know that no one
> ever bothers to adjust those numbers in the field ...

Here's a thought: there are tools (bonnie, ioscan) whose job is
determining details of disk performance. Do we want to look at
creating a small tool/script of our own that would (optionally)
determine the correct parameters for the system it is installed on and
update the appropriate parameters?

--
=====================================================================
| JAVA must have been developed in the wilds of West Virginia. |
| After all, why else would it support only single inheritance?? |
=====================================================================
| Finger geek(at)cmu(dot)edu for my public key. |
=====================================================================