possible computer heat issues

might as well ask about this here too

http://ubuntuforums.org/showthread.php?p...st12398265

How hot are your components when idle on both OS, and when running your games on both OS?

Could be that you're video card gets overheated.
Just download GPU-Z and measure it.

In mint do you have lm-sensors installed? how hot are we talking here?

Unfortunately you can't see the error message you posted unless you're a member of the Ubuntu forums so I'll have to hazard a guess at what it might say.

Dust doesn't directly affect temperature that greatly. Dust getting into fans and causing them to fail does, but good quality fans are generally well sealed and have bearings that go on forever.

As we're talking about a Mobility Radeon, I'm going to guess that this is a laptop and therefore coolers are all very small and are most likely of the blower variety. Laptops frequently aren't put together well so I would try removing CPU and GPU heatsinks, clean them up and the reapply with a high quality silver paste anyway, even if this isn't the problem it should reduce fan speed ups.

Do you notice the fans spinup more than intended? Under Windows I would install Speedfan and see what comes up. Speedfan is one of the tools that there simply isn't a real Linux equivalent too. With Speedfan running, set it logging, push the system with Prime95 and Furmark and see what comes out of it. You will expect big temperature increases but it does depend how far they go. If the temps recorded are within spec then it is not a heat issue.

the error message was the blue screen of death ...

... i'll do proper heat measurements later


   

---

I couldn't figure out lm-sensors ...

my attempt ...

I found kde has a handy little temperature widget built into plasma-desktop ...


  base-temp.jpg (Size: 14.01 KB / Downloads: 181)

I found that may base temperature for my ... cpu? ... is 122 *F (about 50*C)

this is running kde with no effects .. lots of apps open in the background ... but none of them doing intense things ...



  base-temp1.jpg (Size: 31.35 KB / Downloads: 182)

here is the source for the readings[/code]

---

after a quick < 30 sec in xonotic (solarium ... about 50ish inconsistent fps)

the fan jumped up in speed as it usually does ...

temperature reads at 130*F 54*C

As you're getting a BSOD, it almost certainly isn't a GPU overheating problem. Overheating causes hangs or in the case of some graphics cards, artefacts but not BSODs as a BSOD happens within program code, not something the GPU runs.

Looks like lm-sensors is working, otherwise there'd be no temperatures in ksysguard.

As I said, there's still no temperature monitoring software under Linux that's as good as Speedfan, so repeat with Speedfan and set it logging while you run Furmark and Prime95. There can be very quick changes in temperature given different power states that modern CPUs and GPUs run at so hence you need to monitor it, not just see what it is after activity finishes.

Kinda surprised nobody has mentioned running Memtest86. That's usually the first thing I try when dealing with blue screens. Hold shift while starting up your computer to get to the grub menu, select memtest. Let that run for a few hours and see if any errors come up.

heres what memtest dug up ... theres something red on it ... that probably not good

although ... I doubt a bad memory card could possible lead to lower frame rates ... could it?

   

Hard to say what bad memory will do, but it can cause all sorts of wonky problems. You'll definitely want to replace whatever stick is bad.

http://paste.ubuntu.com/1444391/

Speedfan's log from when it blue-screened on my windows boot ...

Although ... im pretty sure its the memory issue at this point

It could be there are a few problems here. I have seen Memtest give false positives before when no system problems were evident but that is unlikely. I would try testing with each stick of RAM separately to isolate the problem. You could also try swapping RAM into different slots (obeying dual/triple channels rules though!) and seeing if it makes a difference. There are occasions when a RAM stick works OK in one slot and not in another.

Now, about this:

62842 102.0 84.0 87.0

Assuming that the thermistors are correctly calibrated, these temperatures are very high. Survivable yes but stable? Maybe not. CPUs have a built in thermal shutdown threshold which might well be 95C for your CPU, I'm sure Intel's specs will list this. GPUs will go higher in general but you shoudl check. That's going a little close, especially as we're now in a northern hemisphere winter. I would dismantle it and reaffix all heatsinks with some good quality thermal paste anyway, even if this isn't related to this specific issue.

Slowdowns in games could be caused by throttling due to high temperatures.

I'll do the memory chips thing when I have time (Wednesday probably)

I'll see what I can do about the heat issues ... I'll see what I can do ... this laptop has always been very hot (and noisy... fan wise) ... I haven't seen a noticeable increase in the heat output though .....

the reason I assumed it was a heat problem at first is because my previous laptop (a f#@$ing hp/compaq) melted the solder on its integrated gpu while on screensaver ... (one of the default ones that came with vista ...) ... and of course it did that while I it was downloading linux (of which I was going to replace windows with ...) .... well ... actually that was the second attempt to get linux onto that machine ... the first attempt the computer's cd drive stopped working (again ... right before I was going make the boot cd ...) and then it took me a couple months to figure out a flash boot ... then it over heated ...

and died ...

that laptop hated me ...



and thats why I went with system 76 ... to get linux on the system first ... then windows later ...

I'll see what I can do about the heat issues ... I'll see what I can do ... this laptop has always been very hot (and noisy... fan wise) ... I haven't seen a noticeable increase in the heat output though .....

You wouldn't expect to see an increase in the heat output. That would stay the same. You might actually feel less heat coming out if it clock throttles due to the high temperatures. Remember that temperature and heat are not the same.

Laptops tend to have annoying high pitched fans and if the noise annoys you it's definitely worth trying to improve heat conduction. Silent PC Review would be well worth reading on this topic.

the fan isn't high piched and annoying ... its just loud ...

when I nedded quiet (which is almost never) I used to be able to throttle cpu though kde's powersave feature ... but recently that seems to have dissapeared in a recent update.

anyways its fixed now (I let my brother do it ... little memory housings are just so delicate ...) ... all he did was take them out ... clean them n64 cartrige style ... and put them back in (although he doesn't know if he put them back in the same slots they came from) anyways ... memory test came clean ... although I'm not done testing to see if everything got fixed

and after a month of running perfectly fine ... its having the same probs again :/

Have you tried repeating the previous trouble shooting steps? Memory test etc.?

Memory is not too fragile to handle unless you are a complete ham fisted oaf. I'm sure you'll be OK doing this.

Run memtest. If it doesn't fail, ignore the RAM. If it fails then isolate the error to an individual memory stick. If it appears on both then it could be some memory controller issue or incorrect speeds/timings. Then clear the dust, reinsert and retest.

If errors persist with the memory then temporarily work with the 1 good stick of RAM. See the the crashing issues you were having still happens. If not, it was the one bad RAM stick and order a replacement.

There are various troubleshooting guides and flowcharts out there on the Internet.

soo ... after 3 blue screens in an hour I opened the computer up myself swapped the memsticks and did some minor cleaning

then I turned on again ... and after 15 minutes ... bluescreen ...

so ... I ran memtest and did something else for a while ... memtest said the computer there were no problems

restarted ... and now its been working fine for the past few hours ...

...

... sigh ..

So maybe not memory. Whatever, next time in bluescreens take out a memory stick and try again. If it still bluescreens then test with the other memory stick. This is about isolating the problem. If you can prove by swapping individual sticks that one stick in particular gives the problem then you know that it is faulty and should replace it.

Even so, it could just be Windows being Windows and BSODing because some software thing is broken.

it might ... might ... have been the flashdrive I was using ...

im not sure ...

nope not the flash drive ... its broke again



f#%@ing windows ... i just want to play descent 3

D:

I'm starting to get under the impression that running memory test actually fixes the problem ...

I'm starting to get under the impression that running memory test actually fixes the problem ...

It won't fix the problem as such, that isn't possible. Is it that after running it and rebooting things are then OK? It could be some odd voltage related stability issue and that on a cold boot you can have problems. Then after doing some activity to warm it up, like Memtest, it then works OK. This is far fetched though/

I really do think you need to take it to someone who knows the machine better like the place where you bought it from. None of your troubleshooting so far has actually pinned down the problem so best to hand it over to someone who can do.