Jump to content

Clevo P150EM BSODs/Kernel Panic - BIOS problem?


cyberdemon

Recommended Posts

Hello Tech Inferno,

I have a P150EM in the form of a "PCSpecialist Vortex IV". I've had it for about a year, and quite often had hard crashes i.e. kernel panics with it in Linux (which I had at first put down to hardware compatibility problems under linux, kernel bugs, etc - the crashes vary in characteristics depending on what kernel i'm running - some kernels are more unstable when idle than when under load, and the crashes tend to be in the power saving functions - update_wait_time, try_to_wake_up etc), but I also occasionally get random BSODs under Windows 7 too (it's a dual-boot setup) - so I suspect I have a hardware problem. :(

I have run memtest86, intel CPU diagnostics, and various other tools but none of them show any issues with the CPU/memory.

However, more recently, when I am running games under Windows (e.g. Fallout 3) I am having a particularly nasty hard crash that makes me suspect the BIOS.

The system freezes (playing the last few milliseconds of audio in a loop) and then gives me a BSOD with very little information - no stack trace, no sys file, just four memory pointers, two of which are zero and the other two I didn't have time to read.

That may be normal for a BSOD, but it is what comes after that makes me suspect the BIOS.

When I reboot the machine, windows will get to the "starting windows" screen with its four coloured orbs, and freeze dead just before it draws the orbs, with the HDD light on continuously. If I then power off the machine by holding the button in, and switch on again, it gives me the ominous error message "Disk read error - press ctrl+alt+delete to reboot". I can reboot a few times, power off / on the machine, and get the same thing - to the point of believing I really have a dead disk. I can even go into the BIOS and see that *both* of my Plextor M5M mSATA drives have disappeared!

However - if I leave the machine off for a while, or if I unplug the power cable and battery, press the power button to discharge it, and then put them back and boot up again, it boots up normally (apart from the 'windows failed to start' screen), and I can check the disk with no errors reported.

This isn't a one-off either - it has happened three or four times in the past month.

This is a particularly spooky kind of crash that whiffs of a BIOS bug.. My question is has anyone else experienced something similar, or am I just overheating my machine?

It has always got quite hot when under full load e.g. playing games, and I wonder if there is any way to clock it down (and perhaps make it quieter too) - certainly the stock BIOS gives no options whatsoever to configure any aspect of the CPU or cooling.. I'm inclined to try Prema's modded bios from these forums, but I don't particularly like the idea of bricking the machine entirely. What do you recommend?

Cheers,

cyberdemon

Link to comment
Share on other sites

Is the cpu overlocked any or at all from factory even though there are no adjustements? Just seems like strange issues. Could be a motherboard problem too not necessarily a bios issue. What speed ram is in the system perhaps its faster than the motherboard can handle. Just throwing some thoughts out as im unsure. Dissapearing drive seems hardware related. Maybe its the ssd itself causing the issues, you could give that a check for errors and check the SMART information.

can you post some temps from a monitoring app for cpu and gpu through some gaming before it happens to see if it is temps..

  • Thumbs Up 1
Link to comment
Share on other sites

Hi mw86 -

I'm not sure what you mean by "overlocked any or at all from factory even though there are no adjustements" - do you mean have the vendor i.e. PCSpecialist overclocked it?

I believe not - certainly I had not paid for that. The CPU is a 2.4GHz i7-4700MQ, which is the slowest of the quad-core chips that were offered. I deliberately avoided the higher-clocked i7s as I didn't want it to unnecessarily eat batteries..

The GPU is a GTX780M however, and that does have a high wattage, but I have not modified it in any way from stock settings. The drivers and BIOS do not offer any configuration for the clock speed for either the CPU or GPU.

You're right it could be a motherboard problem - I have head that some Clevo laptops ship with RAM that needs a higher voltage than the motherboard can supply... I would've thought memtest86 would pick that up though - and I have run that for 24hours with no problems detected. The RAM is: "32GB KINGSTON HYPER-X GENESIS 1600MHz SODIMM DDR3 (4 x 8GB)"

As I say I have already done a full scan on both SSDs and checked SMART - no errors reported there. It would be strange for them both to disappear at the same time without some kind of motherboard or BIOS problem though.

I would be happy to reduce the clock speed (i.e. lower than stock) on either the CPU or memory or GPU if it would make the system any more stable - but as I say, the stock BIOS/drivers don't seem to give me any options whatsoever in that regard. Would the Prima-mod BIOS allow me to change these settings?

I'll see if I can get a log of the temperatures and voltages while gaming and get back to you..

Cheers,

cyberdemon

  • Thumbs Up 1
Link to comment
Share on other sites

Hi mw86 -

I'm not sure what you mean by "overlocked any or at all from factory even though there are no adjustements" - do you mean have the vendor i.e. PCSpecialist overclocked it?

I believe not - certainly I had not paid for that. The CPU is a 2.4GHz i7-4700MQ, which is the slowest of the quad-core chips that were offered. I deliberately avoided the higher-clocked i7s as I didn't want it to unnecessarily eat batteries..

The GPU is a GTX780M however, and that does have a high wattage, but I have not modified it in any way from stock settings. The drivers and BIOS do not offer any configuration for the clock speed for either the CPU or GPU.

You're right it could be a motherboard problem - I have head that some Clevo laptops ship with RAM that needs a higher voltage than the motherboard can supply... I would've thought memtest86 would pick that up though - and I have run that for 24hours with no problems detected. The RAM is: "32GB KINGSTON HYPER-X GENESIS 1600MHz SODIMM DDR3 (4 x 8GB)"

As I say I have already done a full scan on both SSDs and checked SMART - no errors reported there. It would be strange for them both to disappear at the same time without some kind of motherboard or BIOS problem though.

I would be happy to reduce the clock speed (i.e. lower than stock) on either the CPU or memory or GPU if it would make the system any more stable - but as I say, the stock BIOS/drivers don't seem to give me any options whatsoever in that regard. Would the Prima-mod BIOS allow me to change these settings?

I'll see if I can get a log of the temperatures and voltages while gaming and get back to you..

Cheers,

cyberdemon

yeah my typo meant overclocked as you figured. Okay Ill look for your temps buddy and maybe consider the prema bios after posting average temps and seeing how hot it all is getting HWiNFO is great for that ;). It definitely isn't normal behaviour that is happening.

Try downloading resplendence softwares "Who Crashed"

after install open app and hit analyze, please paste the contents here to show more of the bsod's your having could be graphics driver or a driver

it definitely seems like hardware issue being that it happens as you say in windows and linux

  • Thumbs Up 1
Link to comment
Share on other sites

Hi mw86

Below is a screenshot of CPUID's HWMonitor after playing a game for an hour. (Apparently I can't attach anything bigger than 800x600 - can the board admins raise this limit per-user?)

makron.jpg

As I said earlier it does get quite hot, but not ridiculously so - the highest max temp is the GPU which hit 92C, but as I understand it this is quite normal for a GPU under load. The CPU max temp was 81C. AFAIK silicon starts cooking somewhere around 120C.

What might be interesting though is that the voltages fluctuate. "VID" (which I assume is VCore) varies between 0.649V and 1.091V overall. It seems to be higher when the system is under load though.. I assume this is to do with Intel's power saving features.

  • Under ~0% i.e. ambient OS load it flickers between 0.668, 0.678, 0.692 and 0.771, presumably depending on Windows' activity and therefore Intel's power saving.
  • Under 1 thread of heavy load it is pegged solidly at 1.034V.
  • 2 threads it varies between 1.034 and 1.058.
  • at 3 threads it *drops*, and fluctuates between 0.909V and 0.933V, sometimes dipping to 0.885V.
  • 4 threads of load and it is even more variable, fluctuating between 0.885V, 0.909V, 0.883V and occasional dips to 0.860V.
  • Full load (8 threads) it fluctuates between 0.822 and 0.808V, with dips to 0.793V.

Each of these threads I am running is just a simple spin loop e.g. while(1); that just uses one thread at 100%. (it is in fact the command 'dd if=/dev/zero of=/dev/null' in Cygwin)

What is strange is that the voltage drops as the load increases past one thread. Possibly I have a VREG issue? :( Either that or Intel's power saving features are scaling back the CPU speeds/voltages at high multi-threaded loads to stay inside a thermal/power envelope.. :puke:

The system was plugged in during this test, and charging up from a flat battery. I also notice that this laptop charges the battery *extremely slowly* when the system is under heavy load. Again I think this is to do with keeping inside the limits of a power envelope. It is probably not allowed to use more than 190W for the whole system at any one time because the 19V power connector on the back is only rated for 10A. Either the BIOS or the "embedded controller" will be keeping track of this and regulating power budgets accordingly!

I've not tried WhoCrashed yet - sounds interesting.. Not sure it'll have anything meaningful to say about a pure hardware crash though - the BSODs that windows gives tend to be blank. No nice stack trace for WhoCrashed to decode.

As of writing this post I have installed WhoCrashed, but unfortunately I don't have any crash dumps for it to analyse. Apparently Windows can't write crash dumps to disk because I am running without a page file. (32GB should be enough for anybody! :P)

I'll turn the page file back on and post again when I have a crash dump.

Cheers,

cyberdemon

Link to comment
Share on other sites

I also have a second HWMonitor screenshot taken after the above 8-threads load test. (apologies for the double post - my previous post isn't approved yet so I can't edit it)

makron_after_Test.jpg

As you can see, the CPU at maximum load can get as high as 98C. Possibly if I were to run a real stress-test program (as opposed to while(1);) at 8 threads it would go a little higher, but again the Intel power management / embedded controller should step in before it damages itself.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use. We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.