[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20140528151730.GW11096@twins.programming.kicks-ass.net>
Date: Wed, 28 May 2014 17:17:30 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Morten Rasmussen <morten.rasmussen@....com>
Cc: Vincent Guittot <vincent.guittot@...aro.org>,
"mingo@...nel.org" <mingo@...nel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linux@....linux.org.uk" <linux@....linux.org.uk>,
"linux-arm-kernel@...ts.infradead.org"
<linux-arm-kernel@...ts.infradead.org>,
"preeti@...ux.vnet.ibm.com" <preeti@...ux.vnet.ibm.com>,
"efault@....de" <efault@....de>,
"nicolas.pitre@...aro.org" <nicolas.pitre@...aro.org>,
"linaro-kernel@...ts.linaro.org" <linaro-kernel@...ts.linaro.org>,
"daniel.lezcano@...aro.org" <daniel.lezcano@...aro.org>
Subject: Re: [PATCH v2 08/11] sched: get CPU's activity statistic
On Wed, May 28, 2014 at 01:10:01PM +0100, Morten Rasmussen wrote:
> The basic problem is that worst case: sum starting from 0 and period
> already at LOAD_AVG_MAX = 47742, it takes LOAD_AVG_MAX_N = 345 periods
> (ms) for sum to reach 47742.
But it takes only 4 periods (4*32ms) to get ~94% of the way there,
So our half-time is 32 * 1024 ms, now look at \Sum_i=1,4 1/(2^i) =
0.9375. So 4 periods (4*32*1024=131072 ms) gets us ~94% of the way
there.
Now, granted, 4 periods of 32-ish ms is still very long, but nowhere
near as horrid as the full convergence.
Content of type "application/pgp-signature" skipped
Powered by blists - more mailing lists