NVS 160 incorrect tempurature and lowest performance level

terran4000 · 10-24-09, 07:15 AM

Good day everyone,

I have a Latitude E6500 with a NVS 160. With every driver set, newer or older (maybe I didn't try old enough), after a few minutes of constant high GPU usage (ie: video games or CUDA stuff) the performance level drops from the highest/best performance of 3 to the worst: 0.

Of course this results in abysmal performance. The only solution is to turn off the laptop completely and do a cold boot. A mere restart has no effect.

I have tried various options, such as the DWORD's options, Coolbits, and the 'forced peformance level' bash script. All had no effect on this issue.

At first I thought this could be a temperature issue, though it is not. At most, the temperature would reach around 70C (give or take a few) and never approaching 80C. Though, at almost random amount of time (though always a short amount of time) 'nvidia-setting' control panel would start to report the tempurature to be 97C constantly and never ever changing.

This is troubling as that is not the temperature. Using nvclock-gtk as a temperature monitor in conjunction with nvidia-settings I did a small test.

Ran a game and watched the temps. For a while both report the correct temp. Though after a while, the performance drops while nvsettings reports 97C and the 'correct' tempurature, reported by nvclock, was 70C and slowly decreased (due to the decreased performance level).

There are no errors in the logs, no timing errors, no artifacts ... just slowness.

I use Kubuntu 9.04 with compiz off. Testing this in Vista: no slow downs at all ever after several hours of Fallout 3 (fun game btw).

If this slowdown happens in Linux and I reboot into Windows without a shutdown (warm boot), the problem persists in Windows as well.

Only cold reboot works.

Attached is a screenshot of nvidia-settings and nvclock-gtk showing two completely different temperatures ... this is after the problem happened.

Does anyone have any idea if there is a fix for this? Or could it be a hardware issue and should get my laptop fixed?

Thanks.
--PK

terran4000 · 10-28-09, 04:08 PM

Does anyone have any idea how to solve the constant 97C temperature issue?

ledoc · 10-28-09, 05:02 PM

Presumably you tried selecting 'Prefer Max Perfomance' in the PowerMizer panel?

10-24-09, 07:15 AM	#1
terran4000 Registered User Join Date: Nov 2008 Posts: 3	NVS 160 incorrect tempurature and lowest performance level Good day everyone, I have a Latitude E6500 with a NVS 160. With every driver set, newer or older (maybe I didn't try old enough), after a few minutes of constant high GPU usage (ie: video games or CUDA stuff) the performance level drops from the highest/best performance of 3 to the worst: 0. Of course this results in abysmal performance. The only solution is to turn off the laptop completely and do a cold boot. A mere restart has no effect. I have tried various options, such as the DWORD's options, Coolbits, and the 'forced peformance level' bash script. All had no effect on this issue. At first I thought this could be a temperature issue, though it is not. At most, the temperature would reach around 70C (give or take a few) and never approaching 80C. Though, at almost random amount of time (though always a short amount of time) 'nvidia-setting' control panel would start to report the tempurature to be 97C constantly and never ever changing. This is troubling as that is not the temperature. Using nvclock-gtk as a temperature monitor in conjunction with nvidia-settings I did a small test. Ran a game and watched the temps. For a while both report the correct temp. Though after a while, the performance drops while nvsettings reports 97C and the 'correct' tempurature, reported by nvclock, was 70C and slowly decreased (due to the decreased performance level). There are no errors in the logs, no timing errors, no artifacts ... just slowness. I use Kubuntu 9.04 with compiz off. Testing this in Vista: no slow downs at all ever after several hours of Fallout 3 (fun game btw). If this slowdown happens in Linux and I reboot into Windows without a shutdown (warm boot), the problem persists in Windows as well. Only cold reboot works. Attached is a screenshot of nvidia-settings and nvclock-gtk showing two completely different temperatures ... this is after the problem happened. Does anyone have any idea if there is a fix for this? Or could it be a hardware issue and should get my laptop fixed? Thanks. --PK Attached Thumbnails

10-28-09, 04:08 PM	#2
terran4000 Registered User Join Date: Nov 2008 Posts: 3	Re: NVS 160 incorrect tempurature and lowest performance level Does anyone have any idea how to solve the constant 97C temperature issue?

10-28-09, 05:02 PM	#3
ledoc Registered User Join Date: Jan 2008 Posts: 330	Re: NVS 160 incorrect tempurature and lowest performance level Presumably you tried selecting 'Prefer Max Perfomance' in the PowerMizer panel?