Re: CONFIG_NO_HZ_FULL + CONFIG_PREEMPT_RT_FULL = nogo
From: Mike Galbraith
Date: Sun Dec 22 2013 - 00:10:58 EST
On Sun, 2013-12-22 at 05:17 +0100, Mike Galbraith wrote:
> On Sat, 2013-12-21 at 19:21 +0200, Muli Baron wrote:
> > On 21/12/2013 11:11, Mike Galbraith wrote:
> > >
> > > Works, modulo noisy workqueues.
> > >
> > > rtbox:~ # sleep 35 && killall pert& cgexec -g cpuset:rtcpus taskset -c 3 pert 5
> > > [1] 5660
> > > 2400.05 MHZ CPU
> > > perturbation threshold 0.018 usecs.
> > > pert/s: 33 >15.75us: 2 min: 0.04 max: 24.14 avg: 2.51 sum/s: 84us overhead: 0.01%
> > > pert/s: 35 >15.54us: 3 min: 0.04 max: 24.89 avg: 2.39 sum/s: 84us overhead: 0.01%
> > > pert/s: 30 >15.27us: 2 min: 0.04 max: 23.03 avg: 2.64 sum/s: 80us overhead: 0.01%
> > > pert/s: 34 >15.12us: 3 min: 0.04 max: 25.03 avg: 2.51 sum/s: 86us overhead: 0.01%
> > > pert/s: 31 >14.93us: 2 min: 0.04 max: 23.86 avg: 2.60 sum/s: 83us overhead: 0.01%
> > > Terminated
> > >
> >
> > I can confirm this works for me as well, but I have noticed some strange behavior under certain
> > conditions.
>
> Hm, so can I. Same everything, different day, nohz_full is NOT working.
DL980 with same rt7 patchset/config, but with nohz_full=55-63 can take a
while to make up its mind, but does kick in (this boot anyway).
vogelweide:/abuild/mike/:[0]# cgexec -g cpuset:rtcpus taskset -c 60 pert 5
2260.90 MHZ CPU
perturbation threshold 0.024 usecs.
pert/s: 1000 >10.20us: 7 min: 7.21 max: 11.28 avg: 8.00 sum/s: 8000us overhead: 0.80%
pert/s: 600 >10.92us: 5 min: 1.64 max: 26.90 avg: 7.81 sum/s: 4690us overhead: 0.47%
pert/s: 1000 >10.79us: 4 min: 7.08 max: 11.49 avg: 7.86 sum/s: 7863us overhead: 0.79%
pert/s: 1000 >10.70us: 5 min: 7.15 max: 11.90 avg: 7.88 sum/s: 7878us overhead: 0.79%
pert/s: 937 >10.57us: 5 min: 4.30 max: 11.06 avg: 7.81 sum/s: 7324us overhead: 0.73%
pert/s: 1 >9.99us: 0 min: 4.10 max: 5.03 avg: 4.45 sum/s: 4us overhead: 0.00%
pert/s: 1 >9.77us: 1 min: 4.40 max: 10.13 avg: 5.49 sum/s: 7us overhead: 0.00%
pert/s: 2 >14.79us: 3 min: 2.65 max:104.68 avg: 15.21 sum/s: 30us overhead: 0.00%
pert/s: 1 >13.84us: 0 min: 4.31 max: 5.86 avg: 4.82 sum/s: 5us overhead: 0.00%
pert/s: 1 >13.18us: 0 min: 4.03 max: 9.17 avg: 5.19 sum/s: 6us overhead: 0.00%
pert/s: 1 >12.69us: 0 min: 2.83 max: 10.65 avg: 5.85 sum/s: 8us overhead: 0.00%
vogelweide:/abuild/mike/:[0]# cgexec -g cpuset:rtcpus taskset -c 60 pert 5
2260.90 MHZ CPU
perturbation threshold 0.024 usecs.
pert/s: 1 >9.11us: 2 min: 2.87 max: 11.95 avg: 6.12 sum/s: 10us overhead: 0.00%
pert/s: 1 >9.09us: 2 min: 3.09 max: 11.75 avg: 6.04 sum/s: 8us overhead: 0.00%
pert/s: 1 >22.94us: 2 min: 2.73 max:262.06 avg: 33.14 sum/s: 60us overhead: 0.01%
pert/s: 1 >21.53us: 0 min: 2.85 max: 12.02 avg: 5.69 sum/s: 7us overhead: 0.00%
pert/s: 1 >20.26us: 0 min: 3.38 max: 11.94 avg: 5.67 sum/s: 7us overhead: 0.00%
pert/s: 1 >19.12us: 0 min: 2.89 max: 12.11 avg: 5.72 sum/s: 7us overhead: 0.00%
pert/s: 1 >18.05us: 0 min: 3.12 max: 11.28 avg: 5.58 sum/s: 7us overhead: 0.00%
pert/s: 1 >17.24us: 0 min: 2.85 max: 13.20 avg: 6.72 sum/s: 9us overhead: 0.00%
pert/s: 2 >26.85us: 1 min: 2.70 max:201.53 avg: 24.99 sum/s: 50us overhead: 0.00%
pert/s: 1 >25.00us: 0 min: 2.81 max: 11.19 avg: 5.56 sum/s: 7us overhead: 0.00%
pert/s: 1 >23.33us: 0 min: 3.16 max: 11.11 avg: 5.59 sum/s: 7us overhead: 0.00%
pert/s: 1 >21.90us: 0 min: 2.87 max: 11.50 avg: 6.47 sum/s: 10us overhead: 0.00%
pert/s: 1 >20.57us: 0 min: 2.87 max: 11.21 avg: 5.99 sum/s: 8us overhead: 0.00%
pert/s: 1 >19.37us: 0 min: 3.22 max: 11.51 avg: 5.64 sum/s: 7us overhead: 0.00%
pert/s: 1 >18.26us: 0 min: 3.26 max: 10.72 avg: 5.97 sum/s: 8us overhead: 0.00%
pert/s: 1 >27.90us: 1 min: 2.67 max:202.74 avg: 26.52 sum/s: 48us overhead: 0.00%
Hm, size hefty perturbations every 30 seconds (sniff.. eau de vmstat).
--
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/