Karl Cunningham wrote: > Lan Barnes wrote: >> On Sat, April 14, 2007 2:05 pm, DJA wrote: >>> Lan Barnes wrote: >>>> :'-( >>>> >>>> It'll run idle overnight with no problems, but when I load the CPU (top >>>> says 30%), it goes from 38 C to 68 C in about 4 minutes. >>>> >>>> I'm taking it back and making the case that the original diagnosis >>>> charge >>>> should still be in force. >>> Everyone keeps assuming that the problem is related to hardware. It may >>> very well not be. Fan and Processor speed are controlled both at the >>> BIOS level, and by ACPI. Someone mentioned Sensors - Sensors cannot >>> write, it can only read temp and fan settings, and in fact is known to >>> interfere with ACPI (which can and /does/ manage fan/CPU/temp settings) >>> on some laptops. >>> >>> I don't remember if you have Windows running on the laptop, but if you >>> do, then a sure way to see if it's hardware vs. firmware/software, is to >>> see if it overheats in Windows. In fact you may have to install Windows >>> just to get a reliable diagnosis. Looking at hardware only is _not_ >>> going to give a definitive answer unless something is obviously broken. >>> >>> Keep in mind that laptop manufactures' BIOS writers seldom properly >>> follow ACPI and ACPICA (Intel) specs, so many models don't work as >>> expected when it comes to fan/CPU/heat management. In Linux, it's the >>> job of ACPI to compensate. It's a very tough whack-a-mole type problem. >>> But it's getting better fast. >>> >>> There are ways to read your laptop's temp trip points and their >>> associated fan speeds and CPU throttle points. Sometimes just upgrading >>> to a newer kernel can fix the problem. Sometimes just upgrading to a >>> newer kernel causes a problem. >>> >>> I've found it educational and helpful to follow the ACPI4Linux mailing >>> list as well as the one specific to ACPI for my laptop. You may want to >>> do the same. >>> >>> I don't recall what brand of laptop you have, but if you can identify >>> the maker of its motherboard, you may find there is an ACPI kernel >>> module specific to it. For instance, there are for Sony, IBM, Asus, and >>> I think MSI and Toshiba. Do a search in Dmesg or the Syslog and look for >>> any strings containing "ACPI". >>> >> >> Wow! That makes sense. I just got it back after a second day of >> diagnostics. They DID run it under windoze for that and it DID complete a >> night of trouble-free CPU 100% running. They chacked the temp with an >> external thermometer (the laser kind -- gotta get one of those). All was >> well. >> >> Result -- No HW probs that they can see. >> >> I'm sitting here seeing it say it's at 30 C. If I push it, I'm betting >> it'll say 68 C. If it doesn't shut itself off, and if I believe it's >> really 30 - 40, do I care? That's a real question. > > And what's shutting it off when it gets hot? I think you might have > mentioned something in the logs about overheating. Maybe the kernel is > shutting it down because it's reading the wrong temp. >
Suppose there might be any benefit in popping in a live cd (eg, Knoppix) and seeing if behavior is different under the same stress tests? Regards, ..jim -- [EMAIL PROTECTED] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
