LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - Hardware (https://www.linuxquestions.org/questions/linux-hardware-18/)
-   -   Overheating Automatic Shutdown? (https://www.linuxquestions.org/questions/linux-hardware-18/overheating-automatic-shutdown-817414/)

Super TWiT 07-01-2010 09:38 AM

Overheating Automatic Shutdown?
 
I have a question about overheating. I am thinking about running blender on this machine so it will be working sometimes 24/7. I want to make sure that my cooling is up to snuff. Before I go spending money I want to make sure I really *need* any of this stuff. So, I have a few questions.

1. Don't cpus automatically shut themselves down if they overheat? (ie anything newer than a pentium.)

2. What about video cards? Will they shutdown automatically?

I don't really care if I wake up to find my computer shutdown, what is more a priority is that my hardware is safe.

H_TeXMeX_H 07-01-2010 10:34 AM

Modern, new, and good CPUs and GPUs will THROTTLE themselves down if they start overheating. It's different from shutting themselves down, they simply underclock themselves so that they don't overheat.

Whether they will shut themselves down in the case of a sudden, catastrophic loss of cooling, say the heatsink falls off completely, that is a different matter and is up to much debate. I know for a while Intel bragged about its CPUs being able to shut down before meltdown occurs, i.e. the CPU melts and damages the mobo. AMD on the other hand was said to have an inferior method of such shutdown, where there was a chance it could overheat before power was cut to the CPU. This was quite a while ago tho, and the article was clearly biased towards Intel, so I can't be sure if this info is relevant.

Either way, let's assume you don't ever want to put the CPU or GPU in such a position. Many BIOSes have the ability to warn the OS via MCE, in the kernel it is:

Code:

  ┌──────────────────────── Machine Check Exception ────────────────────────┐
  │ CONFIG_X86_MCE:                                                        │ 
  │                                                                        │ 
  │ Machine Check Exception support allows the processor to notify the      │ 
  │ kernel if it detects a problem (e.g. overheating, component failure).  │ 
  │ The action the kernel takes depends on the severity of the problem,    │ 
  │ ranging from a warning message on the console, to halting the machine.  │ 
  │ Your processor must be a Pentium or newer to support this - check the  │ 
  │ flags in /proc/cpuinfo for mce.  Note that some older Pentium systems  │ 
  │ have a design flaw which leads to false MCE events - hence MCE is      │ 
  │ disabled on all P5 processors, unless explicitly enabled with "mce"    │ 
  │ as a boot argument.  Similarly, if MCE is built in and creates a        │ 
  │ problem on some new non-standard machine, you can boot with "nomce"    │ 
  │ to disable it.  MCE support simply ignores non-MCE processors like      │ 
  │ the 386 and 486, so nearly everyone can say Y here.                    │ 
  │                                                                        │ 
  │ Symbol: X86_MCE [=y]                                                    │ 
  │ Prompt: Machine Check Exception                                        │ 
  │  Defined at arch/x86/Kconfig:721                                      │ 
  │  Depends on: !X86_VOYAGER                                              │ 
  │  Location:                                                            │ 
  │    -> Processor type and features                                      │ 
  └─────────────────────────────────────────────────────────────────────────┘

Some BIOSes also have an alarm if the CPU fan goes out, not sure if they have an option of what to do if this happens, check the BIOS anyway.

I say if you get a good, reliable CPU fan, and clean the dust out once in a while, there shouldn't be any problems. Make sure you have plenty of other fans on the case too, and they serve some purpose in cooling. Oh, and use good thermal paste, and only a small amount, just enough to cover the base of the heat sink.

catkin 07-01-2010 01:04 PM

GKrellM (and probably equivalent solutions) can be configured to run a script when settable limits (temperature, fan speed, ...) are passed; that gives a wide range of possibilities! :)


All times are GMT -5. The time now is 02:16 PM.