Re: Disk scheduling policy (Re: NEW_BUFQ_STRATEGY)
Matthew Orgass <[email protected]>
| Newsgroups | gmane.os.netbsd.devel.kernel,gmane.os.netbsd.devel.performance |
|---|---|
| Message-ID | <[email protected]> |
On 2003-12-01 [email protected] wrote: > I can think of a reasonably simple way to tune "N" if we can > characterize the seek time of the disk... but that seems dangerous. The seek time = 0 case at least should be considered since then you actually want to jump around. In particular, now that multi-gigabyte compact flash drives are available (which are usually quite slow), it should be possible to do a background compile and still be able to at least edit and save a text file with minimal delay. On the flip side, DVD+RW can have very high seek times. It seems to me like there should be four or five different types of strategies (or adjustments), with the appropriate one selected per disk (though I know almost nothing about this part of NetBSD...). Matthew Orgass [email protected]