Re: Help understanding slow down

[Jakob Oestergaard]

On Sun, May 23, 2004 at 06:04:55PM -0700, William Lee Irwin III wrote:
> On Sun, May 23, 2004 at 05:57:51PM -0700, Phy Prabab wrote:
> > Just for more clarification, here is a perfect
> > example:
> > 2.6.7-p1:
> > 24.86user 51.77system 2:58.87elapsed 42%CPU
> > (0avgtext+0avgdata 0maxresident)k
> > 0inputs+0outputs (13major+7591686minor)pagefaults
> > 0swaps
> > 2.4.21:
> > 28.68user 34.98system 1:12.34elapsed 87%CPU
> > (0avgtext+0avgdata 0maxresident)k
> > 0inputs+0outputs (5691267major+1130523minor)pagefaults
> > 0swaps
> 
> Thanks. This reveals that the performance regression is system time.

Eh, not if I read the numbers right:

2.6.7-p1: 24.86user 51.77system 2:58.87elapsed 42%CPU

24.86 + 51.77 = 76.63 seconds on CPU, 102.24 seconds of waiting

2.4.21: 28.68user 34.98system 1:12.34elapsed 87%CPU

28.68 + 34.98 = 63.66 seconds on CPU, 8.68 seconds of waiting

So, 2.6.7-p1 spends 16.79 seconds more in the kernel as you observed,
but it spends 93.56 seconds more waiting for I/O (or whatever).

Unless I'm totally missing something, the wait seems to be the
regression.

 / jakob

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/