Propose your solution
Attachments
No attachments.
Duplicates
Comments
|
|
"Every now and then, when my fan is blocked, the computer overheats and crashes"
how does your fan get blocked?
"by warning the user that 'Core temperature is dangerously high'"
Most of the critical hardware components(CPU,GPU, and north bridge) use thermal throttling(dropping the clock multiplier and Voltage) to prevent this from happening. If your system is reaching critical shutdown temperatures(~95*C), you need to service your computer(check the thermal paste/heat sink and fix the fan issues).
|
|
|
tntricker:
True, if the temperature is high the computer could do with a close inspection. My laptop crashed constantly [under load] until I removed the dust that had accumulated under the fans.
And then there are some laptops that are just poorly designed from a thermal management point of view (Zepto comes to mind).
Thermal throttling is inadequate, and I'm not even sure that (mobile) graphic cards even use it.
But that only increases the value of such notification - most users would have no idea that 1) they can monitor the temperature (the applet is not on by default), and that 2) the temperature could be a cause for instability.
If users were warned, much as they are already with poor battery life time, SMART, and other issues, users might prevent issues, instead of simply complaining about 'unstable Ubuntu' when overheats.
|
|
|
I wasn't saying I disagree with the poster's idea, I was just saying if it was common for them then they badly need to maintain their system.
"Thermal throttling is inadequate, and I'm not even sure that (mobile) graphic cards even use it."
Thermal throttling may not always be enough to prevent overload, but if the computer is reaching overload during throttling I don't know if any amount of warning will help a user. By the time they respond it's probably too late and the machine won't have time to shut down anyway.
|
|
|
"I wasn't saying I disagree with the poster's idea,"
Noted.
tntricker:
"I was just saying if it was common for them then they badly need to maintain their system."
True, and most people wouldn't have a clue that they should do something without a notification.
In my case I'm not sure that the system throttled down, it certainly didn't appear that way. It would typically run with 100% CPU load (or heavy GPU-load) for a few minutes before sudden poweroff. If there are others (and I know quite a few) that have ageing laptops then chances are they will be in a similar situation.
The warning would not only help to prevent a shutdown. Assume for instance that the computer is in the process of being dusted down, it's currently at 85C at max load. It would take another year of use before it reaches shutdown-temp of 105C. It's still too hot, but the user wouldn't know, except if the user is of the inquisitive kind.
|
|
|
"It's still too hot, but the user wouldn't know, except if the user is of the inquisitive kind."
I guess it wouldn't hurt to report critical temperatures even to an inquisitive user.
|
|
|
|
My computer is critical all the time :P i dont need the warning.
|
|
|
This is a very interesting idea.
But should have a option do unable this feature. In case of a "mistake"
|
|
lolobu
wrote on the 8 Feb 10 at 21:27
|
|
|
|
From time to time, the fan on my PC doesn't start. So I did something like that myself. I've added in the root crontab a script that runs every minute and if the computer is over a chosen temperature, it just shutdown and log the event.
|
|
|
|
Just throwing this out there: Most of the computers I've looked at seem to misreport the tempature or the wrong max tempature is set. If either of these are off, well, any response to an "issue" will either come too late, not at all, or even when there is no issue at all.
|
|
|
"The BeOS programmer's guide covers two functions IsComputerOn (returns 1.0 if computer is on, unspecified otherwise) and IsComputerOnFire (returns temperature if mainboard has flames coming from it, unspecified otherwise). It's right there in the printed version (though I quote from memory)."
.. it might be relevant ;)
|
|
|
|
And if a sensor is "fucked" up, your computer will always shutdown? I got a sensor like that, it says 128C and always 128C but it is not true.
|
|
glococo
wrote on the 16 Feb 10 at 18:44
|
|
|
You are talking about a propietary nvidia driver.
Nvidia can use the Bubble Notification to show warnings, but unfortunally from the opensource community, we cant do anything with nvidia driver. Its closed.
Maybe the ATI driver will do that.
Regarding the Motherboard & CPU sensors, there are lots of fake temperature reports from builtin sensors. My motherboard report rightnow 75º but I know that its at only 45º.
Anyway, maybe we can use the bubble notification system or a NEW way like "bad sectors warnings" :)
|
|
Areso
wrote on the 19 Feb 10 at 22:15
|
|
|
The screenshot dialog descriptions are reundant.
It should be more like:
Browne Internet [Firefox]
Burn CD [Brasero]
Play Music [Amarok]
Scan documents [gScan2pdf]
...
It is enough to mention the program one time.
|
|
|
|
An OCD notification is nice, but that's a terrible way to handle a potentially melting situation. We presume that the motherboard has safety measurements to automatically turn off the PC before it melts ... i dunno.
|
|
krs
wrote on the 1 Mar 10 at 16:37
|
|
|
This is a bad idea. Every processor and every graphic card have a specific critical temperature.
Unless we have access to an up-to-date data-base with all processors, chipsets and gfx cards, such a tool will mostly provide false-positive warning.
An overheating PC is a bad designed PC, or a broken PC, it's a hardware issue, and it's not the purpose of a software to fix a hardware problem.
|
Post your comment
|
