On Sat, Feb 23, 2008 at 01:31:00PM +0100, Andi Kleen wrote:
> > *) compute the context-switch pair time average for the system. This is
> > your time threshold (CSt).
>
> This is not a uniform time. Consider the difference between
> context switch on the same hyperthread, context switch between cores
> on a die, context switch between sockets, context switch between
> distant numa nodes. You could have several orders of magnitude
> between all those.
Wouldn't the common case for blocking on a lock with a short hold
time be the minimal context-switch pair on the same CPU?
> > *) For each lock, maintain an average hold-time (AHt) statistic (I am
> > assuming this can be done cheaply...perhaps not).
>
> That would assume that the hold times are very uniform. But what happens
> when you e.g. have a workload where 50% of the lock aquisitions are short
> and 30% are long?
>
> I'm a little sceptical of such "too clever" algorithms.
Me too. That said, I cannot resist pointing out that the measurement
of interest would be the fraction of lock-hold times that are shorter
than the context-switch time. ;-)
Thanx, Paul
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
