On Fri, Jul 26, 2002 at 10:45:23AM -0600, Steven Cole wrote:
> I found that dbench gives significantly lower numbers when
> the partition on which it is run is mounted as ext3.
Yes, especially with ext3 using its default 5-second commit intervals
and ext2 using 30 second timeouts for its own bdflush.
dbench does a lot of IO to temporary files, writing large amounts of
data and then shortly deleting it. If it can run fast enough to
delete the data before it hits disk, it goes _much_ faster, so you end
up measuring cache speed and not disk speed. ext3 is particulary hurt
by that given its shorted default commit timeouts.
Cheers,
Stephen
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/