Hi,<br>
Is there any documentation on how to use powermgr?<br><br><div><span class="gmail_quote">On 7/27/05, <b class="gmail_sendername">Thomas Tuttle</b> <<a href="mailto:tom@...">tom@...
</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Okay, the best way to do that would be to (using powermgr) create a<br>
profile named "hot" and then write your script like so:<br><br>#!/bin/pseudocode-interpreter<br>if (TEMP > $HOT) {<br> echo "System is actually hot, slowing everything down.";<br> powermgr --profile hot;
<br> sleep 5;<br> if (TEMP > $HOT) {<br> echo
"System is about to explode, shutting down.";<br> poweroff;<br> } else {<br> echo
"System has cooled down, returning to normal."<br> powermgr --auto;<br> }<br>} else {<br> echo "System is just being foolish and is not hot.";<br>}<br><br>The "hot" profile would probably look like this:
<br><br>Profile hot<br>cpu frequency = 0%<br>EndProfile<br><br>That way it would slow down the CPU frequency without playing with other<br>stuff like brightness.<br><br>On Wed, 2005-07-27 at 10:33 -0400, Devon Miller wrote:
<br>> Thanks Thomas, I'll definitely check it out. One feature I would like<br>> to see is the ability to throttle the cpu to manage temperature.<br>><br>> My laptop occasionally reports spurious high temperatures. Within 3
<br>> polls the temperature might be reported as 70C, 97C, 70C. When the<br>> kernel sees the 97C it powers off the system.<br>><br>> I've hacked around this by patching the kernel to run /sbin/overheat<br>> instead of /sbin/poweroff. Overheat checks the temp again and if it's
<br>> still hot, shuts down powernowd and sets the cpufreq to its minimum<br>> value It then sleeps for 5 seconds and if the temp is still 90C+,<br>> calls poweroff.<br>><br>> The downside is the system is now left in a very slow state. I have
<br>> not written something to bring it back to a dynamic clocking state. A<br>> daemon that would manage all of this would be really appreciated! (Of<br>> course, the kernel would still need to be patched to not poweroff
<br>> until the daemon has had a chance to try cooling things down.)<br>><br>> <dcm><br>><br>><br>> On 7/13/05, Thomas Tuttle <<a href="mailto:tom@...">tom@...
</a>> wrote:<br>> I've been working on a program called powermgr. It's a daemon<br>> written<br>> in Perl that can control many power management functions on<br>> Linux,<br>
> including CPU frequency and/or governor, screen brightness,<br>> laptop mode,<br>> fan speed, wireless power management, as well as runlevel and<br>> services,<br>> based on the state of the system.
<br>><br><br><br>-----BEGIN PGP SIGNATURE-----<br>Version: GnuPG v1.4.1-ecc0.1.6 (GNU/Linux)<br><br>iD8DBQBC556Z/UG6u69REsYRAo8qAJ9w8ab8s183nvc3cN/AXh1tqeUt3wCfYSUT<br>fu3lk34QqS3XwoUdpQtK48U=<br>=/Lqt<br>-----END PGP SIGNATURE-----
<br><br><br></blockquote></div><br>
|