M860TU / GTX260M Problem (Random freezes in games)

Hi guys,

I would like you to help me with this desperate cause. I have a Clevo M860TU with Q9000, GTX260M, 4GB DDR3 RAM etc. for 3 years now. It served me perfectly for 3 years, until last month. Since then i've been experiencing a problem.

I start a game, play for 4-5 minutes with no problems but then, my screen freezes, goes black for 10 seconds and comes back to game again. When it happened, it told me that display driver crashed and recovered successfully. This started to become so frequent that i can't play games anymore. Also, whenever it occurs 3 times, game completely freezes, making me have to CTRL + ALT + DEL and kill the process.

To me, it was an obvious problem of GPU. Since i clean my laptop's fan regularly, i thought it was the thermal-paste that gone corrupt. So a few days ago, i bought a new thermal paste and applied it. On top of that, i reinstalled windows 7 x64 to make sure the problem would be gone forever.

Here comes my dissapointent, the freaking freeze still happens! I checked my GPU temperatures and it was 70 C when freezing happened(CPU 60-65), so i don't think it's a heat problem anymore. Google'd my problem, looked everywhere, there are similar problems, some people claim they solved it by doing 'X', some claim they couldn't solve it. I tried all of 'X' like reverting back to 2009 drivers, etc. but none worked. I tried 186.81(this was stated as the most stable driver), 25x.xx, 27x.xx, 301.24, 306.23 but no cure.

Back to start, i got no solution at all. What could be wrong with my PC ?

Thanks in advance for your answers...

 

welcome to nbr
those temperatures are fine.
if youve tried 186.81 and they failed then your buggered as they are the most stable drivers for the 8/9/200 cards.
the other option would be to try some drivers with modified inf from laptopvideo2go but ive never used mods so i have no idea how to modify them.
next we need to check everything else. could be hard drive,could be a dodgy ram chip etc.
download mem test and HD Tune and run some tests on your ram and hard drive and report back.

i had a similar scenario a few months back. i wasnt gaming but i was getting black screens on boot up and very slow and slugish and then freeze ups on just normal brousing which turned out to be 1 bad sector on my WD scorpio black.

 

Thanks for your fast reply. I wil try your offers and report back now, i hope this isue will be solved...

 

forgot to mention you can get all the software from my sig below.

 

The same problems crop up on the desktop 200 series cards as well, I was discussing this in a different thread in the owners lounges. From what I've researched there is no fix, so basically buggered like MrDJ said. I've heard of a script being able to fix it but any attempt to procure has led me nowhere. It's a known issue with Nvidia but they haven't listed it as a real high priority goal to fix unfortunately. My problems started occuring when I switched from 32bit to 64bit if you did something similar or wanted to try out a 32bit environment to see if it works.

 

I am afraid it sounds like video card is defective.

You can go to Unigine: real-time 3D engine (game, simulation, visualization and VR) and download the testing program for video.

 

As i said in my last post, i tried checking my HDD and RAM also. HDD shows 2 bad sectors, could they make a problem like this ? I will try to repair them. But now i need to leave to go to school now, when i come back i will run RAM test and last 2 posts' recommendations. Thanks for replies, i will report when they are done.

 

ohh yes. i had only one bad sector of my original wd scorpio black 320gb and it seriously caused some very annoying problems.
boot up would just hang. black screen after full boot and hard turn off to reboot. then bsod, and then once it did boot lockup as soon as i clicked on a shortcut.
all happened about 3-4 months ago and it took 4 total clean installs before it worked. then changed hard drive and it was sorted. still slower than before and sluggish at times but it was a 4 year old lappy.

 

Okay now i repaired the bad sectors by using the windows' own tool. HD Tune shows no bad sector. It took so much time that i can just start RAM tests. Meanwhile i tested my GPU with furmark, it passed the benchmark and successfully did the burn-in test up to 90 C with no problem at all. But whenever i start a game(spesifically Heroes of Newerth), even at loading page black screen problems keeps happening... Skyrim is fine for 3-4 minutes, then black screen happens 3 times and voila, crash. Now HDD is okay, GPU has no temp. problems. Only RAM left as a hardware, so now i am on to that. Let's see what memtest will say but i think i lost my hope and started to think that my laptop is old now. Will be back after memtest, thanks for your interest again.

 

Good morning. I did RAM test and had no error on that. Then a strange thing happened. I underclocked my GPU from 550/850/1350 to 400/600/1000 and i had no black screen, just some heavy games were slow and that should be because of underclocking. And i tested my notebook with OCCT and it did the same black screen while testing GPU. It created some graphics, i can upload them later today. They essentially show that fps drops to 0 sometimes, and immediately goes back to what it was, hence the black screens. Now i have started to think it's rather a GPU's memory problem(since occt does same blackscreens on vide memory testing) or a PSU problem, that my GPU can't draw enough power at full speed. I will test underclocking gradually when i come back, and see what are the safe numbers close to stock. Hope you can help me with my new theories, i will report back when i complete things i mentioned tonight. Have a nice day

 

i will have to leave this one to someone else as ive never overclocked or underclocked so wouldnt know where to start.

 

That's very interesting. You've gone beyond what I personally have tested but I look forward to hearing the results.

 

One of the strangest things i've ever seen on a computer, it works very stable at 500/750/1300 mhz, instead of 550/800/1350. I played skyrim for 2 hours with no problem at high settings, seems like problem is somehow solved, i didn't notice any performance drop with these new clocks. So it's enough for me, thanks for your help guys

 

Thanks for sharing, it didn't occur to me to give the underclock a shot but I'm glad it worked for you. Hope you continue to enjoy your system!

 

Sorry for reviving this thread but I'm having a similar problem and this for the past 2 months. Basically my issue is way more random but when it happes there is no other option than restarting the laptop (The screen turns black and sound stutters in the background). I did the standard tests: temps,drivers,complete windows reinstall, then running only with the most important drivers and still the same problem.I also checked the memory for about 10hours with memtest86+ I did a burn in test with FurMark 20minutes and all ok. Harddisc diag tool also passed.

I would like to give this underclocking a try, any Ideas how to do this?

Thanks,
Onne

 

I do the underclocking with the own tool of nvidia. You can use this link,

http://us.download.nvidia.com/Windows/nForce/system_tools/6.05/6.05_nvidia_system_tools.exe

Try and report please, if it helps you too, i will make sure it's related to GPU's clock speeds. I have no problems right now, except some performance loss due to underclocking.

Will wait for your reply, hope it gets solved.

 

Thanks for your reply, I will do a OCCT run first to see if I get the same result than you. I will move on to the underclocking sometime this weekend as I have other things to do aswell .
Did you do a default test or a customized? By this I mean if you changed the values for: FPS Limit, Shader Complexity, Resolution, FullScreen, Error Check, etc for the GPU test.

UPDATE: I managed to reproduce my error with the GPU test of OCCT, however I couldn't generate a log because the crash completely freezes the notebook and apparently the log files get generated once the test finishes.
I will mess around with OCCT a bit more and go with the underclocking later today or tomorrow.

 

Firstly, i am sorry for late reply, could hardly find time this week. I did the default test as well, but my crashes didnt freeze the computer, so i could get the log file. But still, i think having it reproduced implies that your problem is GPU, too. So i wonder the underclocking results now, hope it gets solved soon

 

Hello I just now gave the underclocking a try since I have been really busy lately. I'm running a GF 9800m GTS 512mb default vaules are: 600 Mhz Core clk, 800 Mhz Mem clk, 1500 Shader clk.

I'm running the card @ 550, 750, 1400 Mhz atm and testing. Since the crashes happen randomly and a few days ago OCCT ran for 30min wthout a crash at stock values it will take some time for me to report back and to determine if the underclocking is beeing succesful.

 

How is it going mate, any improvements ?

 

Made an account just to answer this. Was having the same problem with the GTX 260M on my Gateway FX P7908u. Suddenly last week it started crashing during games. Scoured Google 'till I came across this thread. Underclocking to mert469's settings works for me. Problem solved.

 

I am so glad that i could help you Have fun with your machine, mate !

 

Hi sorry for me reporting so late but I have been quite busy during the past 10 days, writing lab reports and other stuff for university and haven't really had time to test.

I underclocked my PC about 1 week ago down from: 600/800/1500 Mhz (GPU/Memory/Shader) to 550/750/1400 Mhz. Yesterday it crashed once while playing league of legends. I'm going for a underclock to 500/700/1300 Mhz and see how it goes. So far OCCT:GPU ran for 1:30h without problems but from my experience in games it crashes randomly after 20min launching the game or after having played 5-7 gaming hours.

The freeze is basically a full system freeze the follwing happens:

- If Im listening to music it starts stuttering
- the external USB keyboard becomes unresponsive and if i press "Numlock" the corresponding key on the keyboard doesn't light up.
- Screen turns black (but the backlight is still lit)
- alt+ctrl+del doesn't work.
- fans still spin
- upon restart, there is no entry on the event viewer, only (the system has restarted without clearly shutting down).

What I already have done:

- I'm not overclocking.
- Notebook adapter has been changed to a more powerful one (to rule out power probs).
- BIOS up to date, has been for more than a 1 year now and it has been running great for atleast 8-10 Months.
- Windows complete reinstall and update to all the latest patches incl. SP 1.
- Running the latest Clevo drivers.
- Tried running the Latest Drivers from Intel (mobo,etc.) aswell.
- Heat is optimal: CPU @ 55°C tops under load, GPU hits 74°C only when doing OCCT:GPU test else it runs at 69°C ish.
- Memtest86++ Ran for 12hours.
- HDD diagnosis tool via DOS.
- Never crashed during desktop apps.
- OCCT: CPU ran without problems.
- managed to crash it once running OCCT:GPU 3D test.
- Tried another HDD.

I once was looking for laptop cooling methods and came across this video: Custom Made HDD Laptop Water Cooler - Video
I played it and my PC crashed with the same symptoms. Youtube however works fine.

UPDATE: 11.Nov.2012 - So far I have had about 4-5gaming hours among the past 10 days and so far no crash. It appears that the Notebook is running more stable. Ill g et back once I hit the 12 hour mark and after a more extensive gaming session.

 

Quick heads up:

- Notebook still freezes, and it happened about 3 times during desktop apps aswell, like browsing my files on the hard drive while listening to music so nothing intensive. The notebook GPU was running at underclocked speeds. I have run out of ideas ...
-Memory has been retested for 18h this time with memtest and everything passed. (Total memory test time counting my previous tests go to about 30h +)
- I have tried 3 different Hard Discs problem still happens.
- Yesterday I tried the 3rd HDD and it frozes while performing the "winsat.exe" process to rate the performance of my system afte I installed the nvidia drivers shortly after having done a clean Windows install.
- Have done a Intel CPU diangnosis with Intels tool downloaded from their site, CPU is running like 60°C under the maximum allowed temp and everything passed.
- My last option is to start dissabling components on my motherboard, but I really doubt that such a "hard" crash is getting caused by a mafunctioning driver (note there is no BSOD nor event viewer entry).

I guess I didn't have luck on this one and I might have to look for another option to make this Notebook hold a bit longer till I can upgrade to a newer one.

Onne

UPDATE: I reverted to the stock clock and had 6 crashes on desktop operations during the last 2 days something never before happened. I noticed that a bunch of nTune profiles were getting loaded once logged into windows. Next to my custom (downclock) profile was another (default nvidia) profile that I suppose manages the dynamic clock adjustments (common in notebooks) to save power.
So my theory for the moment is that my old GPU has problems adjusting the dynamics clocks and crashing the system. Anyway I did a further downclock to 450/650/1250 Mhz and removed the "dynamic clock adjustment" profile.
Since the GPU is not worth changing (replacement is about 250 Euros) for a 4year old system, I just need to push the notebook for some more months till i can upgrade.

Ill report back.

 

UPDATE as of Jan 08 2013:
With my Downclock to 450/650/1250 Mhz I only had 1 crash during the last 2 months and this while having an external monitor Plugged and gaming Full screen on that monitor, while using the Notebook Monitor for facebook and such. So far I have been unplugging my external monitor for gaming and I has been running great, so I guess I have overcome my problem as for now.

I do realize however that my VGA is most likely damaged, and it might get worse at some point. I will be trying to spot the exact clock that is causing the issue during the next weeks.

I hope this thread helps anyone else encountering similar problems.

~Onne

 

You could try upping the core voltage a touch, but if this is degradation setting in it will only speed the issue.

 

Onne, how did you manage to downclock your GPU?
I think I have a similar problem as yours.

 

The link was posted on by the original poster on page 2. Reposting the link here:
us.download.nvidia.com/Windows/nForce/system_tools/6.05/6.05_nvidia_system_tools.exe

The tool is quite straightforward. Note that uppon restart the clocks get reset to the default values unless you create a profile under the nvidia tool to load your custom values uppon windows logon.
Anyway I posted a list of pretty much everything I have done until attempting the downclock on one of my posts, maybe you can get some ideas of what else to try.

Good luck

 

UPDATE: 05. Feb 2013.

I managed to reproduce the crashes really quickly with Futuremark 3D Vantage, (you can find it on google, free dl). I started tweaking the independent clocks (GPU,Memory and Shadder) and realized that 3D Vantage doesn't crash when running the GPU and Shadder Clock at stock speeds, while the Memory clock is underclocked. So atelast I for now don't have to underclock the entire Card.

This has also led me to think that it might be temperature related after all, as the readings from GPU-Z / HW-Monitor are only the reading from the GPU-Core. It might just aswell be that the VRAM chips and related components on the card are running to hot and this can't be tracked with the thermal sensors, which originally made me discard overheating as the GPU Sensor was showing a max of 70°C ish under load.

Anyway, I hope this new info helps other people experiencing the same or similar problem. I'm going to take the notebook appart and reaply thermal compound to the VRAM Chips and all related stuff I can think of.

~Onne

 

You may find the information in this article useful:
Clevo M860TU / Sager-NP8662 / mySN XMG5 GPU (GTX260M -> FX-3700M Replacement / Upgrade and Temperature Management Modifications/

Nvidia GPUs compe pre-overclocked from the factory, often past stability limits. As the silicon ages through use and abuse, the stable limits drop.

It is generally a good idea to test stability with OCCT, and retune the clocks voltages until you have 100% stability after at least a couple of hours of OCCT testing and the temperatures stay in a sane range (ideally < 85C after a few hours of OCCT testing).