[ltp] Temperature increase after S2Ram

Henrique de Moraes Holschuh linux-thinkpad@linux-thinkpad.org
Mon, 8 Jul 2013 00:07:16 -0300


On Sun, 07 Jul 2013, Paul Bolle wrote:
> On Sun, 2013-07-07 at 02:47 -0300, Henrique de Moraes Holschuh wrote:
> > This is not the first report that 3.10 is a dud.  Make extremely sure you
> > have _NOT_ enabled the new "full tickless" mode, that thing is broken to
> > kingdon come [either that, or it gets all zombies to come out in the open].
> 
> I currently use
>     $ grep NO_HZ /boot/config-3.10.0-1.local0.fc18.x86_64 
>     CONFIG_NO_HZ_COMMON=y
>     CONFIG_NO_HZ_IDLE=y
>     # CONFIG_NO_HZ_FULL is not set
>     CONFIG_NO_HZ=y
>     # CONFIG_RCU_FAST_NO_HZ is not set
> 
> So the new full tickless mode is disabled, isn't it?
> 
> > Check with powertop
> 
> Somehow it seems powertop triggers Heisenbug behavior! If the second
> core is apparently stuck at 2 GHz, and I start powertop to see what's
> going on, the second core will again drop down to 800 MHz, eventually.
> But quitting powertop makes the second core jump to 2 GHz again (judging
> by /proc/cpuinfo's output, that is).
> 
> That is strange! So let's assume for now that I see patterns that just
> aren't there. Ie, this needs more testing on my side.

Nah, let's assume 3.10 is borked.  Please check whether the latest 3.9.x
works well for you...

> > and turbostat
>     $ sudo ./turbostat -S
>     No invariant TSC

I see.  Can powertop show idle states on that box?  If it does, keep an eye
on them.

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh