Bad crash and I'm not sure why

I have a G73J and just now when I was running GTA IV my computer crashed, no BSOD the screen just went black and then the computer shut off. What was really terrifying was that it wouldn't turn back on. I hit the power and the lights on the keyboard just flicked on for a second then off. I tried it again and nothing at all. I unplugged everything, took out the battery, and waited for a minute. It finally turned back on and everything was working fine, no error messages or anything acknowledging the crash. What I'm really concerned about is that it was heat related. I live in the desert and there is a tone of dust everywhere. When it was off I blew into the back and a ton of dust came flying out.

I ran SpeedFan and got about 50-55 C when not much is running. During a benchmark it got to about 70 at the highest. That's hot but it seems in line with what I know about these computers. So do you think what happened was heat or dust related or could it have been a software issue? If so why didn't the computer turn back on immediately? I'm running Catalyst 10.7, I think I had tried to revert it to 10.1 at some point and I'm not sure if that's just not displaying correctly or I accidentally let ASUS live update change something. Would going to 11.3 help? Also does anyone have any ideas about combating pervasive dust?

Sorry if this reads as freaking out a bit but I live at the end of the earth and if this computer blows I'm going to lose my mind.

 

Run HWINFO32 and check each section of the GPU temps under full load for 10-15 minutes and see what you register.

Personally that sounds like a thermal shutdown and the G73 will not turn straight back on again until it has reached a safe temperature.

If it is temp related blow compressed air through the vents to shift any dust.

Have you updated the VBIOS? Repaste might be an idea if all else fails.

 

I'm not sure I 100% understand HWINFO32 and I didn't have a benchmark thing that would run for 15 minutes, was more like 5, but the GPU numbers seems to run between 65 C when not doing anything to 85 C at 100% But I don't know what's normal for this.

 

Get furmark and run a stability test. 1280x1024 with xtreme burning, post FX and 4x MSAA is enough to push your GPU to 100%. Load HWiNFO32 the temps sensors for your GPU are where it says either nVidia or ATI, would be great if you told us if it's a JH, or JW too.

Start furmark and monitor your temps closely, if there's something overheating you'll probably see it in 10 minutes. When my JH was overheating, i would hit over 100C in 5 mins on furmark.

 

Sorry I missed a letter before I have a JH. I ran Furmark with the settings you suggested. It was at 99% GPU and hit 100 C in just over 5 minutes. So is this the problem? How hot should it run with the GPU working that hard? Also is there something I can do to prevent it from overheating to the point where it does permanent damage?

 

Permanent damage probably would start around 105C.

You should be worried. You can either read the NUMEROUS threads started explaining how to redo the thermals on your graphics card. Or you can bring it to a laptop repair shop and they can do it for you. You can call Asus and ask to find out where the nearest authorized service center is for you.

I would use ICD7, OZC Extreme, or MX-4. Those are all good and easily available.

Notice the title on this sticky? Would help to do what it says.
http://forum.notebookreview.com/asu...blems-useful-threads-read-before-posting.html

 

I can't do any of this any time soon. I live in the middle of nowhere China. Is there some program that can just get i to cut off if it gets too hot? Or some other idea?

 

Does it go above 100oC or settle around 100oC? Thats pretty much ok to be honest for the JH because remember your pushing it to way past normal operating conditions with Furmark and that should not have been the reason GTA caused it to shut down as thermal shutdown can be anywhere between 105oC to 114oC and most games run 5-15oC lower than Furmark and GTA is very CPU intensive.

Both fans running according to Furmark? most importantly the CPU fan?

Get a can of compressed air and short sharp bursts into the vents to start with could be clogged with dust. Treat a repaste as the last option because ambient temp reflects harshly on a GPU's running temp and you might not see a major difference.

 

I'd imagine majority of the thermal pastes are made in China. So not being able to do repaste while in China has to be a joke.

 

According to AMD, permanent damage starts at 114 i think. Anyways those temps are high, you should expect around 90-95 max on furmark. Clean the vents first, a lot of people don't do that and it's the source of their heat problems.

Like Dallers said ambient temps impact a lot. On 22.5C, i get 79 max with my last repaste, i would probably get a few more degrees lower at 21C amb. If it's like 30C ambient where you are then i wouldn't worry. Clean your heatsinks no matter what though it's good for the computer in general. That's one of the reasons my toshiba M40 is still running today (bless that copper heatsink too).

 

Well it's got to be at least partially heat related. I updated my drivers and ran GTA IV for a while checking HWINFO32 for the temp, but it never went past 88. I tried Furmark again though and crashed after about 2 minutes, as soon as the screen went black I shut down rather than wait for the computer to shut down. I rebooted, waited a few minutes and tried again, with the same result. The temperature got into the low 90's and then it shut down. Furmark didn't record the fan as working, it only listed one, but HWINFO32 did and I could hear it kick into high when the temp got above about 80. I really can't get thermal paste, or even compressed air, I live in the dead middle of nowhere in a desert. I'll probably just avoid any game where this happens as from everything I've run so far GTA IV is the only thing giving me this issue. Unless anyone has some other suggestion or explanation. Thanks for all the help.

 

We seem to be focusing on the GPU here what are your CPU temps?? you should have two fans listed by HWINFO32 if only one fan is listed and your CPU is not, the fan could have packed up or is not working. While running check the left fan as if your facing the screen is blowing. This wont cause an immediate shutdown unless stressed hard as I7's are made to be pushed to the limit.

If not we are back at stage one. When you say live in a desert? Is it sandy much? Blow hard into the rear fans when its turned off.

 

I ran Furmark again on somewhat more moderate settings. For some reason the line in HWINFO32 where it shows what percent of the GPU is being used isn't showing up anymore. The GPU got up to 95-6 after about 5 minutes but the first core only got to 70 and stayed there. As the GPU got into the mid 90's the screen flicker a little and it actually went black for a brief second before coming back to normal, where I stopped the test. Also just to note HWINFO32 lists 4 GPU temperatures:

GPU Thermal Diode
GPU TS0 (DispIO)
GPU TS1 (MemIO)
GPU TS2 (Shader)

The first and third got to 96 the fourth got to 91 but the second only got to 78.

Finally there is definitely some air coming from the left side of the computer. When the GPU gets really hot and I hear the fan rev up more air comes out the right side but some is coming from left also.

 

As above ^^ CPU what are your CPU temps!? The GPU temps are fine and would not cause a thermal shutdown.

Just restart HWINFO32 with sensor only if it stops working.

 

The CPU temps for that test maxed out at 70, 59, 63, and 68.

Edit: The amount of GPU used it still missing.

 

Hey there, I know your pain. I had a similar issue with my Asus G53JW and it is similar to yours. I wonder if the reason is the same:

check this

 

Just to get more info, though I'm not sure it will help.
What bios are you using?
What vBios are you using?
When you reinstalled the drivers, did you uninstall the older ones first then used driversweeper 2.9.0?
What drivers are you using, cat 11.3?

Get the latest version of HWiNFO32 which is 3.73. There's a lot of readings and they don't all fit in the default sensor window. Also Get GPU-Z while you're at it, this one will show GPU utilization for sure.

When you say your GPU got up to 96 in 5 mins, does it still go up afterwards if you let it run for more than 5 minutes?

Your CPU temps seem fine by the way. You'd want to run a prime95 torture test to check at 100% CPU usage though, furmark doesn't really stress the CPU, only the GPU.

 

Actually, it can operate up to 114C safely (according to AMD), however, Asus set Thermal Shutdown at 110C to play it safe.

 

OK after some more tests it's definitely something to do with the GPU and heat. I ran Prime95 for 10 minutes and the hottest CPU core never got above 75 C. But when I ran Furmark, and for some reason in both HWiNFO32 (newest version) and GPU-Z I can't get the GPU load to read anything but 0 no matter how hard it's working, when the heat go into the low 90's thing started getting odd. The screen blinked a few times and in one test it just shut off for a second at around 95. Also in one test the cursor started to look weird and it didn't go back to normal until I restarted it.

In the first test I did I got to 100 C with no problems but now I'm getting all sorts of weird things when I approach that temperature. Since I can make it happen in Furmark, and always when it gets hot it seems that the GPU just doesn't work well with that much heat.

Edit:

One more odd thing. I was running HWiNFO32 in the background while playing some other games, just to see what sort of heat I was getting, and I noticed the GPU Utilization number was working again. I ran another test in Furmark at 1280x720 with 4xMSAA and Dynamic Background on and it read as 99.9 on the GPU Utilization. The heat got up to 90 C pretty quick but then it just stay at around 92-93 C for the whole 15 minute test I ran. That's much much longer than any other test I ran and there was no problem. I don't know if that's just a fluke or if all these tests and what not knocked some dust loose in there, but it's good news at least. I don't want to run too many more tests though since I'm worried about doing damage that way.

Thanks again to everyone for taking the time to help me with this problem.

 

A complete shutdown during game = overheating

I had this problem 3 weeks ago (laptop hitting 110C during games) and I thought I had to repaste, but I did this and it lowered my temperatures by 35C:

1. unplug and remove battery
2. set laptop upside-down on table
3. remove the bottom access panel
4. use some canned air and spray in short bursts into the laptop through the panel you just removed. line the shots up with the fans.
5. do the same process except blow through each fan vent and towards the access panel
6. if you see any amount of dust, you probably just fixed your problem
7. keep doing it until you see no more dust

8. if that doesn't work, check these boards on re-pasting instructions. you can buy the Diamond thermal paste for $7 on various computer supply sites