Re: re-inline sched functions

[Nick Piggin]

Andrew Morton wrote:

> "Chen, Kenneth W" <kenneth.w.chen@xxxxxxxxx> wrote:
>
> > This could be part of the unknown 2% performance regression with
> > db transaction processing benchmark.
> >
> > The four functions in the following patch use to be inline.  They
> > are un-inlined since 2.6.7.
> >
> > We measured that by re-inline them back on 2.6.9, it improves performance
> > for db transaction processing benchmark, +0.2% (on real hardware :-)

Can you also inline requeue_task? No performance gain expected, but
it is just a simple wrapper around a list function.

> > The cost is certainly larger kernel size, cost 928 bytes on x86, and
> > 2728 bytes on ia64.  But certainly worth the money for enterprise
> > customer since they improve performance on enterprise workload.
>
> Less that 1k on x86 versus >2k on ia64.  No wonder those things have such
> big caches ;)
>
>
> > ...
> > Possible we can introduce them back?
>
> OK by me.

What happens if you leave task_timeslice out of line? It isn't exactly
huge, but it is called from a handful of places.


-
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/