On Sun, 30 Mar 1997, Erik E. Fair wrote:

> Further, given that full understanding of the knobs in FFS is limited, I
> wonder:  are there any tools that one can run on a raw disk (or on a
> filesystem) that can exercise the disk in some way, possibly analyze the
> filesystem if one is present, and then suggest potentially better FFS
> parameters, and project the likely performance increase?
> 
> Or are FFS's existing defaults already giving us 90% of the theoretical
> maximum performace for the general case (and thus such a diagnostic or
> exercise program wouldn't really be worth writing)?

In my experience, the biggest performance hindrance for modern systems and
disks is the default block and fragment sizes. I usually change the default
8K/1K to 32K/8K. Real life performance is vastly improved and the
difference in speed between reading the raw disk device and reading a large
test file is not noticeable, as opposed to the normal 1:3 ratio. The
fragmentation wastage for "normal" filesystems seems insignificant, too.

This is with modern SCSI disks on ahc adapters.


-- 
Soren