I did a small test today to try test the speed impact of
the KSE test.
during this, I noticed that soft updates is not having it's usual
performance increaseing effect..
Normal kernel, soft updates:
1387.881u 603.392s 35:48.63 92.6% 2694+2248k 12424+3310io 3587pf+0w
KSE kernel, soft updates:
1385.979u 605.575s 35:21.41 93.8% 2695+2252k 11558+3300io 3705pf+0w
Normal kernel, NO softupdates:
1389.803u 594.961s 35:59.14 91.9% 2697+2248k 11839+66940io 3701pf+0w
It's interesting that despite the the extra writes (63640 of them)
it took almost the same time from all metrics.
this is a 35 minute subsection of buildworld..
i.e. buildworld up to a particular point,
there may not be anything wrong but I expected more improvement.
I remember that when I first committed softupdates it made a noticable
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message