Re: patch: p4tcc and speedstep cpufreq drivers

> Date: Thu, 24 Feb 2005 13:11:14 -0800
> From: Nate Lawson <nate_at_root.org>
> 
> Maxim Sobolev wrote:
> > Kevin Oberman wrote:
> >> No joy. I set it to 262 and it was fine. The next step killed the system
> >> again.
> >>
> >> I'm also concerned that taking TCC out of automatic mode might not be a
> >> great idea, at least until things like _PSV are supported. When I do a
> >> buildkernel, buildworld or any big compile job, I need to slow down the
> >> CPU to keep the CPU form frying. It quickly jumps to 185 F. or higher if
> >> I don't. If I understand automatic TCC, it should throttle the CPU all
> >> by itself to prevent this.
> > 
> > 
> > Taking TCC out of automatic mode doesn't disable thermal controlling 
> > circuitry completely, so that if the processor overheats it will shut 
> > down the machine anyway:
> > 
> > ---
> > Regardless of enabling of the automatic
> > or On-Demand modes, in the event of a catastrophic cooling failure, the 
> > processor will
> > automatically shut down when the silicon has reached a temperature of 
> > approximately
> > 135 °C. At this point the system bus signal THERMTRIP# will go active 
> > and stay active
> > until RESET# has been initiated. 
> > ---
> 
> Correct.  Even more so, automatic mode continues to override On-Demand 
> mode if there is a more moderate thermal condition than THERMTRIP#:
> 
> "On-Demand mode may be used at the same time Automatic mode is enabled, 
> however, if the system tries to enable the TCC via On-Demand mode at the 
> same time automatic mode is enabled AND a high temperature condition 
> exists, the duty cycle of the automatic mode will override the duty 
> cycle selected by the On-Demand mode."
> 
> Since automatic mode is set by the BIOS before we even boot, things 
> should be fine.
> 
> >> Between throttling and frequency adjustment I can get about 16
> >> performance levels and I don' see a good reason for another 15. Also,
> >> the change is frequency is so non-linear that small changes often don't
> >> make sense. The first three step are fairly straight, but then things
> >> get bumpy. It looks to me like all frequency settings are not created
> >> equal. 
> > 
> > I wonder this too. I think in the presence of several independent 
> > regulators we need some form of calibration to get more or less precise 
> > results.
> 
> You can manually test this kind of stuff by doing:
> hint.p4tcc.0.disabled="1"

Thanks! With this set I am down to "only" 14 frequencies, but they are
now proportional to CPU speed and I still have a fully functional TCC,
should I need it.

Also, at no added charge, the system no longer locks up at low speed. I
can to all the way down to 150 MHz just fine! No more hangs.

I really think that if would be a good idea to have
hint.p4tcc.0.disabled set by default, especially based on sobomax's
recent message where the spec says that normal operation can't be
assured if the TCC is not in automatic mode.
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman_at_es.net			Phone: +1 510 486-8634