MX250 crashing

Peer

Well-Known Member
Joined
May 17, 2018
Messages
587
Reaction score
305
Credits
402
Hey guys,
after some time I got my nvidia MX250 in my laptop up running with optirun. And it works fine. Accept that after some time for example playing minecraft my laptop just crashes. I tried a bit around and found, that the thermals are a bit high using the gpu. So I opened up my laptop and removed some dust from the cooler (not much). The laptop is not very old.
Now games run just fine, I played some minecraft and some xonotic and had no issues. Then I used a shader in minecraft (heavy gpu load) and it crashed again. I also ran a benchmark wich crashed after some minutes.
I don't know what to do except reapplying thermal paste.
But that should not be the issue, because the laptop is very new.
Does someone know how to fix this.
BTW I use arch.

Thanks
 


Even though the laptop "is very new" it does sound as if the problem is a thermal issue.
The thermal paste or thermal pads could have shrunk, dried up, or been improperly applied by the manufacturer.

Whereas cleaning out some dust did seem to help I would seriously consider replacing the thermal paste and/or thermal pads.
 
Hey guys,
after some time I got my nvidia MX250 in my laptop up running with optirun. And it works fine. Accept that after some time for example playing minecraft my laptop just crashes. I tried a bit around and found, that the thermals are a bit high using the gpu. So I opened up my laptop and removed some dust from the cooler (not much). The laptop is not very old.
Now games run just fine, I played some minecraft and some xonotic and had no issues. Then I used a shader in minecraft (heavy gpu load) and it crashed again. I also ran a benchmark wich crashed after some minutes.
I don't know what to do except reapplying thermal paste.
But that should not be the issue, because the laptop is very new.
Does someone know how to fix this.
BTW I use arch.

Thanks
You may open ur laptop again and remove the old thermal paste and apply a new thermal paste and then ON ur laptop press ur BIOS key then reinstall the Operating System completely by deleting all the partition again in such process the data in all the drives will clear and reinstall all ur apps by updating the apps
I think it will work
Thanks You
 
You may open ur laptop again and remove the old thermal paste and apply a new thermal paste and then ON ur laptop press ur BIOS key then reinstall the Operating System completely by deleting all the partition again in such process the data in all the drives will clear and reinstall all ur apps by updating the apps
I think it will work
Thanks You
I don't think so, because this also happens when booting from an external drive with Ubuntu. And the problem are the thermals I checked the and they where an around 90 so °C.
 
then reinstall the Operating System completely
Hey @Peer, unless you see a clear indication that something went wrong on the software side of things by browsing the system logs, or unless the system keeps crashing after repasting, you do not need to reinstall anything.
 
Even though the laptop "is very new" it does sound as if the problem is a thermal issue.
The thermal paste or thermal pads could have shrunk, dried up, or been improperly applied by the manufacturer.

Whereas cleaning out some dust did seem to help I would seriously consider replacing the thermal paste and/or thermal pads.
Thanks for both of your replys.
It looks like, I have to go nad buy some thermal paste. But due to corona I may not go to the city very soon.
 
Good luck! That is something I never had to do yet... but I guess I'll enjoy the process when I do.
 
I've got bad news. I'm afraid under Linux the thermal throttling on many Nvidia/Intel laptops is pretty useless. The trend is for the BIOS to only provide protection by shutdown and hand over complex thermal control to the OS using ACPI cnfiguration. Unfortunately ACPI is so complicated and messed up that most suppliers only do the minimum to get it working with Windows. Poor old Linux has to try and pick up the threads and make sense of it all and doesn't always do a good job. I've had three laptops from HP, Dell and Toshiba with similar problems. The HP caught fire!
Some games run OK, it's all about how hard they drive the GPU and how sustained that is. SuperTuxKart seems to be particularly bad for example, similarly GPU accelerated rendering in Blender.
Having said that, there are applications (e.g. thermald for Intel machines) which do a pretty decent job at throttling the processor; but the GPU is in the closed world of Nvidia. If you look at Nvidia XServer settings under the thermal tab you will see the thermal slowdown temperature. On my machine it is set to 97 degrees which is far too high. The processor shuts the machine down at 100 deg C before the GPU can throttle. If I tweak the CPU power limits in thermald (which requires PhD level meddling) the CPU will slow to about 400MHz but the GPU still sometimes runs away and cooks the machine. The MX250 does not let you change the slowdown temperature through nvidia-smi, it's not clear where it is set.
So I am stuck. maybe one day a BIOS update or new Nvidia driver will solve this, but I'm not optimistic. Maybe someone else out there has an answer. Browsing the internet seems to suggest that thermal paste will win you 5degC or so at the risk of destroying your machine. A periodic dust removal is vital, your decision whether to delve in the case to do this. If I had been better at dusting my HP would not have set it's voltage regulator on fire. Next time buy the least sexy laptop with the biggest case and noisiest fan.
 
I've got bad news. I'm afraid under Linux the thermal throttling on many Nvidia/Intel laptops is pretty useless. The trend is for the BIOS to only provide protection by shutdown and hand over complex thermal control to the OS using ACPI cnfiguration. Unfortunately ACPI is so complicated and messed up that most suppliers only do the minimum to get it working with Windows. Poor old Linux has to try and pick up the threads and make sense of it all and doesn't always do a good job. I've had three laptops from HP, Dell and Toshiba with similar problems. The HP caught fire!
Some games run OK, it's all about how hard they drive the GPU and how sustained that is. SuperTuxKart seems to be particularly bad for example, similarly GPU accelerated rendering in Blender.
Having said that, there are applications (e.g. thermald for Intel machines) which do a pretty decent job at throttling the processor; but the GPU is in the closed world of Nvidia. If you look at Nvidia XServer settings under the thermal tab you will see the thermal slowdown temperature. On my machine it is set to 97 degrees which is far too high. The processor shuts the machine down at 100 deg C before the GPU can throttle. If I tweak the CPU power limits in thermald (which requires PhD level meddling) the CPU will slow to about 400MHz but the GPU still sometimes runs away and cooks the machine. The MX250 does not let you change the slowdown temperature through nvidia-smi, it's not clear where it is set.
So I am stuck. maybe one day a BIOS update or new Nvidia driver will solve this, but I'm not optimistic. Maybe someone else out there has an answer. Browsing the internet seems to suggest that thermal paste will win you 5degC or so at the risk of destroying your machine. A periodic dust removal is vital, your decision whether to delve in the case to do this. If I had been better at dusting my HP would not have set it's voltage regulator on fire. Next time buy the least sexy laptop with the biggest case and noisiest fan.
I had thought of throttling the GPU but figured that doing so would interfere with the game play.
Perhaps it depends on the game being played and how much throttling can be done before degrading the experience.
 
Which distro are you running? Where did you obtain optirun?
 
Arch like I mentioned in my first post.
I read the arch wiki article.
 

Members online


Latest posts

Top