Question Driver power state failure

[oliveria]

hello good people, I've experienced this trouble for quite a long period of time with my dell precision 7720, of driver power state failure, I've read online about how on how to solve it but still not getting it right, like today I've installed a new OS and I've disabled windows from installing drivers through updates, right now the only two drivers I've installed are the intel hd 630 intel and Nvidia graphics drivers P3000 from Nvidia website, I already downloaded these two drivers and they were in my flash drive so I haven't connected the pc to internet after installing windows, what surprises me is that its running into the driver power state failure every time even when its just idling, I haven't done anything apart from installing those two drivers from my pen drive, prior to this I even tried the DDU uninstaller but the problem is just too persistent, at times when I click on restart it takes like 2 minutes to initiate the restart process, Will gladly appreciate your answers

 

[Aeacus]

Hmm....

intel hd 630 intel and Nvidia graphics drivers P3000

Disable the Quadro P3000 from the BIOS. Look if you still get the same BSoD error as usual.

If not, issue is with your Quadro and/or it's drivers.
If issues remain even when running only iGPU, issue is far more severe.

Since what you have is a laptop, component replacement isn't an option.
If you would have desktop workstation, things would be FAR easier.

 

[oliveria]

Hmm....


Disable the Quadro P3000 from the BIOS. Look if you still get the same BSoD error as usual.

If not, issue is with your Quadro and/or it's drivers.
If issues remain even when running only iGPU, issue is far more severe.

Since what you have is a laptop, component replacement isn't an option.
If you would have desktop workstation, things would be FAR easier.

when I disable Quadro from device manager, I don't run into this blue screen of driver power state failure, is it possible to download older version of Nvidia graphics drivers, I tend to know an older version that used to work with this laptop and wanted to try it, 8s it possible to get it from Nvidia website

 

[Aeacus]

is it possible to download older version of Nvidia graphics drivers, I tend to know an older version that used to work with this laptop and wanted to try it, 8s it possible to get it from Nvidia website

Yes and Yes.

1. Go here: https://www.nvidia.com/en-us/drivers/
2. Use the "Manual driver search". E.g put in P3000 to the search bar.
3. It then shows plethora of different OS types.
4. Pick the correct OS you have on your laptop.
5. It then auto-fills the options.
6. Click "Find".
7. Site thinks a bit and redirects you to the results page.
8. Latest GPU driver is visible, but click on the "View More Versions".
9. It then shows all the GPU drivers that are compatible with your GPU, ordered by newest at the top.
10. Once you've located the correct driver version you're after, click on the "View" just behind it, on the right side.
11. Site then redirects you to download page, where you can download the drivers with Download button.

Once drivers are downloaded;
1. DDU old drivers.
2. Reboot.
3. Install new drivers.
4. Reboot.
5. Look if things got better.

 

[oliveria]

Yes and Yes.

1. Go here: https://www.nvidia.com/en-us/drivers/
2. Use the "Manual driver search". E.g put in P3000 to the search bar.
3. It then shows plethora of different OS types.
4. Pick the correct OS you have on your laptop.
5. It then auto-fills the options.
6. Click "Find".
7. Site thinks a bit and redirects you to the results page.
8. Latest GPU driver is visible, but click on the "View More Versions".
9. It then shows all the GPU drivers that are compatible with your GPU, ordered by newest at the top.
10. Once you've located the correct driver version you're after, click on the "View" just behind it, on the right side.
11. Site then redirects you to download page, where you can download the drivers with Download button.

Once drivers are downloaded;
1. DDU old drivers.
2. Reboot.
3. Install new drivers.
4. Reboot.
5. Look if things got better.

Great and Thanks for the reply, I've done exactly this and installed an older version of Nvidia that works before, but now I'm getting another new blue screen called video scheduler internal error, or could this means my graphics card is failing and I need to get another mobile workstation, if I could get a confirmation that it's my graphics card that is failing, I would just get another laptop to remove me from this stress, each time I install my Autodesk architectural softwares to start working then the blue screens pop up and I cant just work in peace, ive installed windows several times I get the blue screens and I tend to think it could be a failing graphics card, what other idea could you be having please

 

[Aeacus]

but now I'm getting another new blue screen called video scheduler internal error

This error means many things, including incompatible (old) GPU drivers and also GPU hardware failure as well. There are other things it may mean, but for you use case, with clean Win installation, other aspects doesn't apply.

You could try version or two newer and see if it helps (excluding latest version which is confirmed not to work). But other than that, there's nothing else to do.

or could this means my graphics card is failing and I need to get another mobile workstation

I'd look towards new PC. Desktop PC would be better since you can replace individual components when need be. (E.g if your current PC would be desktop, it can be easily fixed with GPU replacement). But if you need portability, then laptop is only option.

Also, Intel HD630 means 7th gen Kaby Lake CPU. That architecture is now almost 8 years old and new, better (beefier) PC would be in order.
Sadly, nothing lasts for forever and we all need to update our hardware from time to time (either when it gets obsolete and slow or breaks and doesn't work right anymore).

 

[oliveria]

This error means many things, including incompatible (old) GPU drivers and also GPU hardware failure as well. There are other things it may mean, but for you use case, with clean Win installation, other aspects doesn't apply.

You could try version or two newer and see if it helps (excluding latest version which is confirmed not to work). But other than that, there's nothing else to do.


I'd look towards new PC. Desktop PC would be better since you can replace individual components when need be. (E.g if your current PC would be desktop, it can be easily fixed with GPU replacement). But if you need portability, then laptop is only option.

Also, Intel HD630 means 7th gen Kaby Lake CPU. That architecture is now almost 8 years old and new, better (beefier) PC would be in order.
Sadly, nothing lasts for forever and we all need to update our hardware from time to time (either when it gets obsolete and slow or breaks and doesn't work right anymore).

Yes I truly agree with you, I've also run out of hopes that this problem could be driver related, I've tried two older graphics drivers but all haven't worked and still just getting various different blue screens, at one time its video tdr failure, another time its video management internal error, another time its page are in non paged areas, another time its video scheduler internal error, and all these errors I see in one way or the other to be related to the graphics card hardware, I think I'll just plan for another PC, I bought this 1.5 years ago on ebay and for all this while it had been working just well, but I remember it has been producing those single beep sound which I read are related to an underlying trouble either in the motherboard or in the graphics card, I could hear such single beep sounds occasionally like even when adding volume, a sound sounding as if sb has dropped a tiny piece of dust on a keyboard, last month it made that sound once while I was typing and completely shutdown and didn't turn on till I had to take it to a technician, ever since I came back from the technician, its been blue screens, and I cant do anything with it, I'll just find another laptop, and I really appreciate for the information you've given me

 

[oliveria]

Hello, I ran a windows diagnostics to find out which hardware error my pc could be having, Im pasting the hardware report here so you can help me identify which specific hardware is this that could be the culprit, I was not able to analyse it and get it well, thanks


Description
A problem with your hardware caused Windows to stop working correctly.

Problem signature
Problem Event Name: LiveKernelEvent
Code: 141
Parameter 1: ffffa484b90ce460
Parameter 2: fffff80749c616a0
Parameter 3: 0
Parameter 4: 29e8
OS version: 10_0_19045
Service Pack: 0_0
Product: 256_1
OS Version: 10.0.19045.2.0.0.256.48
Locale ID: 1033

Files that help describe the problem
WATCHDOG-20241222-2206.dmp
sysdata.xml
WERInternalMetadata.xml
WERInternalRequest.xml
memory.csv
sysinfo.txt

View a temporary copy of these files
Warning: If a virus or other security threat caused the problem, opening a copy of the files could harm your computer.

 

[Aeacus]

Code: 141

That's hardware failure code. Can also mean corrupted OS.

To remove the 2nd possibility;

1. Open Start.

2. Search for Command Prompt (or cmd.exe), right-click the top/correct result, and select the Run as administrator option.

3. Type the following command to perform system repair and press Enter:
SFC /scannow

4. If the scan found corrupt files and repaired them, run the same command again, just in case:
(if no errors were found, skip to step #5)
SFC /scannow

5. Once no errors are found with SFC /scannow, type the following command to perform a quick check and press Enter:
DISM /Online /Cleanup-Image /CheckHealth

6. Once this is done and when corruption is detected, type in another command for in-depth scan and press Enter:
DISM /Online /Cleanup-Image /ScanHealth

7. Once this is done and there are issues with system image, type in another command for fixing it and press Enter:
DISM /Online /Cleanup-Image /RestoreHealth

8. Once the system image repair is complete, type in another command and press Enter:
SFC /scannow

9. After last once completes, final command with Enter press would be:
exit

This closes the terminal window.

Note: SFC /scannow, DISM /ScanHealth and DISM /RestoreHealth may take a long time to complete. Do not close the terminal window or do not reboot the system, when each of theses scans/fixes are taking place, even when they seemingly get stuck.

 

[oliveria]

I've done all these steps as you've outlined, but the driver power state failure is still occuring frequently, I've even tried using one ram, swapping them, to no avail, the pc came with Samsung ssd 850 pro sata, and I bought an nvme on ebay which I've installed the OS on, I want to try installing the new OS back to the old sata ssd 850 pro and see if I'll still be getting the blue screens, when it came with the OS on the 850 pro ssd, its never had these blue screens, I guess this could be my final try and if the blue screens will be there then I think this is a dead machine

 

[oliveria]

after installing a new OS on the former Samsung ssd 850 pro that came with the laptop , and installing the intel hd 630 and Nvidia graphics drivers from respectve websites, I'm now getting an error code 9f in the windows event viewer after crashing, I've now ran out of options, before I decide on starting to get another laptop could there be anything more to add as advice,

 

[oliveria]

also what surprises me is the fact that all these blue screen troubles I've only experienced them with dell precision mobile workstations, I previously owned an hp zbook 17 G3 workstation, and generally other hps I've owned haven't given such stress and time wasting blue screens, in my next step I'm planning to drop the dells and try again other hp, I could be wrong still

 

[Aeacus]

error code 9f

This error code means that a driver is not handling the power state properly. Either due to the incompatible drivers or due to the out-of-date drivers.

When GPU starts acting up, it may not recognize once compatible drivers anymore, spewing errors left and right.

in my next step I'm planning to drop the dells and try again other hp, I could be wrong still

Sure, feel free to go with HP (or Lenovo) when your experience with Dell has been bad.

You do not have to stick with one brand. Since we have brand selection out there, feel free to go with another brand.

Even i have brands that i avoid since i've had poor experience with them.

Spoiler: My brand experiences, optional to read (click here to view)

One such brand is Adata. I'll never buy any Adata storage media ever again.

Another brand that i avoid is Gigabyte. Now, i had Gigabyte monitor for years (namely their subsidiary brand TopView) and for the TN panel, it was decent monitor, albeit it had brightness issues (whole panel looked way too dark). And i had Gigabyte GZ-G2 Plus PC case as well, while looking fancy on the outside, in actuality was terrible to work in (e.g cable routing, fan mounting). And from my profession, i've come to learn that Gigabyte loves to do shenanigans with their MoBos and MoBo revisions. So, i'm not getting any Gigabyte product for myself anymore and won't be suggesting it to others as well.

Sometimes, one has to change the trusted brand not because there have been bad experience with it, but because the latest/current offer is worse than that of a competitor.

I've used Intel CPUs since 4th gen Haswell architecture times. While my current rig is using 6th gen Skylake architecture CPU (one generation older than the one in your laptop). And now, i'm on the look out for upgrading my CPU, MoBo and RAM as well.
I waited until Intel 15th gen Arrow Lake architecture released (Q3 2024), just to see how it fares against 14th gen Raptor Lake architecture. And with new architecture, Intel also renamed their CPUs. Used to be, where CPUs were named simple, e.g: i5-6600K, i7-7700K, i3-8100, i5-9500, i5-1600K, i7-11700K, i9-12900K, i3-13100, i5-14500.

Whereby: Core I3/i5/i7/i9 showed entry-level/mid-range/high-end/top-end.
Then: the 1st number after Core i shows generation. E.g i5-6xxxK is 6th gen, i7-11xxxK is 11th gen, i5-14xxx is 14th gen.
Last three numbers show the CPU within the segment, e.g: i5-6400 is weaker than i5-6500. But i5-6500 is weaker than i5-6600.
And last letter suffix meant different things based on the letter. E.g K-suffix = CPU has unlocked multiplier and can be OCd. F-suffix = CPU doesn't have iGPU in it.

But latest 15th gen Intel CPUs are all renamed to Core Ultra and the naming doesn't make much sense either. In 15th gen lineup, we have:
Core Ultra 5 245K, Core Ultra 5 245KF, Core Ultra 7 265K, Core Ultra 7 265KF and Core Ultra 9 285K.
5x CPUs and that's it. In actuality, 3x CPUs since the KF-variant is same CPU but without iGPU in it. Also, no Core i3 equivalent, which would've been good for office PC builds.

And the Core Ultra reviews showed clearly that there are almost 0 improvement over the predecessor 14th gen CPUs in terms of performance. While Core Ultra CPUs cost considerably more.
I was eyeing Core Ultra 5 245K as my new CPU, but it fares very poorly. Only thing it has going for it, is lower power draw. But i plan to put it into desktop PC and i have 80+ Titanium efficiency PSU to power it (the most efficient PSU there is), so, i don't care about CPU power draw. Also, most of the times, i don't do heavy CPU utilization tasks.
So, i looked towards AMD and their latest, namely Ryzen 5 9600X, which is equivalent in terms of the segment where CPU is (mid-range). But Ryzen 5 9600X beats Core Ultra 5 245K in all tasks, both gaming and application. On top of that Core Ultra 5 245K CPU costs far more than Ryzen 5 9600X. With the price of Core Ultra 5 245K, i can get Ryzen 7 9700X, which is huge step up in terms of performance.

And this is where i'm at. I don't dislike Intel. I did hope that the new Core Ultra would be far better than it is. I could've even soak up the higher price, IF the performance would be there. But it isn't. Core Ultra 5 245K costs the same as Ryzen 7 9700X, while being far inferior in terms of performance. Thus, it is no brainer to go with AMD CPU.

Sorry for the long rant but i just wanted to tell how even long preferred brands can be abandoned due to the poor products they are releasing.

At current date, none of the Intel 15gen CPUs can be suggested in the desktop PC segment, since AMD Ryzen 9000-series CPUs are both cheaper and better performing than the latest Intel CPUs.

Though, where Intel is doing good now, is in GPU segment. Namely their Battlemage (B-series) GPUs. Intel ARC B580 (MSRP $249) and Intel ARX B570 (MSRP $219), which are priced very competitively and would be great for any user who uses 1080p monitor.

 

[oliveria]

Great, thanks for the advice, and as a last resort do you think undervolting this Quadro P3000 is possible, in other videos online I did read other guys managed to solve the video tdr failure through undervolting

 

[oliveria]

each time I tru new Nvidia graphics driver installation I just run into blue screens and in the windows problem reports I just see a persistent hardware error code 141, so at least for now I can be convinced that its my GPU that has failed

 

[oliveria]

and its worth noting that in each Nvidia driver I was trying, I was going through the process of DDU to remove the old drivers before trying others, I tried the top three new ones and some 3 older ones but nothing worked

 

[Aeacus]

do you think undervolting this Quadro P3000 is possible

GPU itself is already low power draw GPU, only 75W. So, there isn't much to undervolt.

Undervolt isn't a magical thing to make GPUs work better. In actuality, you'll risking GPU stability to gain lower GPU temperatures and less power draw.
And even if you could undervolt a GPU, many GPUs won't let the undervolt to stay. Instead, GPU itself will override the undervolt you apply and will work as nothing happened.

In one of the other topics, i described in-depth what GPU undervolt is compared to setting a power limit for GPU. There, the GPU in question was RTX 4070 Ti but same applies to all GPUs:

Power limit is different from undervolting.

In undervolting, you are changing the available voltage level of a GPU core.
RTX 4070 Ti Super default core frequency is from 2340 Mhz at ~835mV. As the frequency goes up, the core voltage follows to sustain said frequency. With turbo frequency of up to to 2610 Mhz at ~925mV.

In undervolting, you lower the voltage for the core on all frequencies. E.g when default voltage range for 2340 Mhz to 2610 Mhz was 835mV to 925mV, then undervolt would bring the voltage range to 825mV to 915mV (-10mV offset).

With this, and with less voltage to the vcore, GPU doesn't get that hot. But GPU may not be stable either. It could be stable in the frequency range of 2350 Mhz to 2500 Mhz but once GPU goes higher than e.g 2500 Mhz, the lowered vcore voltage isn't enough to keep the GPU stable and GPU crashes, crashing entire system.


Power limit limits the wattage GPU gets. RTX 4070 Ti Super 285W GPU (it actually can consume more, but for simplicity sake, let's say 285W is max what GPU pulls).

With no power limit, GPU is freely able to pull as much wattage it needs, up to 285W. While following the default frequency/voltage cure it has (2340 Mhz at ~835mV to 2610 Mhz at ~925mV).
But if you limit the power, e.g to 250W, GPU frequency increases until power limit is met. E.g up to 2550 Mhz at ~912mV.

GPU core will always be stable, since it gets appropriate voltage at any given frequency it runs at, but you just limit on how high the GPU can boost it's core frequency. But with undervolting, you try to starve the vcore at any given frequency, for lower temperatures.


It's complex to explain it and i'm not sure if you understood it; so, perhaps analog example makes more sense;
Cars (everyone knows cars). GPU power limit is like the speed limit on the road. If speed limit on the road is up to 90km/h, you can drive up to 90km/h, while your car consumes 10L per 100km. While without speed limit, you can drive as fast as you want, e.g 150km/h, while your car still consumes 10L per 100km.
While undervolting is like, rather than providing the car engine 10L of fuel per 100km, you limit it to 8L per 100 km. With this, you'll gain fuel efficiency but you are starving the car's engine of the fuel it needs. And that at every speed you drive at, be it 30km/h, 90km/h or 150km/h. And with this, you can actually seize up the engine. Or when it comes to GPUs, crash the GPU and PC.

 

[stonecarver]

and its worth noting that in each Nvidia driver I was trying, I was going through the process of DDU to remove the old drivers before trying others, I tried the top three new ones and some 3 older ones but nothing worked

Honestly I think Aeacus has you on the right track. For the sake of getting it out of the way have you gone to Dell and installed it's official Nvidia drivers.

Support für Precision 7720 | Treiber und Downloads | Dell Deutschland

NVIDIA Quadro Pxx00/Txx00/Mxx00/RTX xx00/RTX Ax000/RTX Axx00 and GeForce RTX x0x0 Graphics Driver and NVIDIA Control Panel Application

Critical​

Video

12 Nov 2024

 

[oliveria]

Honestly I think Aeacus has you on the right track. For the sake of getting it out of the way have you gone to Dell and installed it's official Nvidia drivers.

Support für Precision 7720 | Treiber und Downloads | Dell Deutschland

NVIDIA Quadro Pxx00/Txx00/Mxx00/RTX xx00/RTX Ax000/RTX Axx00 and GeForce RTX x0x0 Graphics Driver and NVIDIA Control Panel Application

Critical​

Video

12 Nov 2024

This

Honestly I think Aeacus has you on the right track. For the sake of getting it out of the way have you gone to Dell and installed it's official Nvidia drivers.

Support für Precision 7720 | Treiber und Downloads | Dell Deutschland

NVIDIA Quadro Pxx00/Txx00/Mxx00/RTX xx00/RTX Ax000/RTX Axx00 and GeForce RTX x0x0 Graphics Driver and NVIDIA Control Panel Application

Critical​

Video

12 Nov 2024

Yes its true Aecus has done a great job, I want to try this driver from dell though I already tried it before but still got the blue screens, if I look at the Nvidia website, their latest release of quaro p3000 is dated 10th Dec 2024 and this from dell is 12th Nov 2024, why do you think Dell takes longer posting the relatively newer drivers, I'll try it now and see

 

[oliveria]

GPU itself is already low power draw GPU, only 75W. So, there isn't much to undervolt.

Undervolt isn't a magical thing to make GPUs work better. In actuality, you'll risking GPU stability to gain lower GPU temperatures and less power draw.
And even if you could undervolt a GPU, many GPUs won't let the undervolt to stay. Instead, GPU itself will override the undervolt you apply and will work as nothing happened.

In one of the other topics, i described in-depth what GPU undervolt is compared to setting a power limit for GPU. There, the GPU in question was RTX 4070 Ti but same applies to all GPUs:

Ive gone through the undervolting you've described here, its a very well detailed explanation, I tried to used msi afterburner before to try to even reduce the core clock speed by 10 percent, but I didn't succeed, could there be any other software or alternative around that can tweak anything in this P3000 by even 10 percent to see if things can be peaceful with me, I'm just trying anything that can give me some slight hopes to be able to use this pc to complete a pending task before I get another one in one months time, or could it be that my GPU connect to the motherboard is loosely held or something could be a miss in that connection, I'm even thinking of unscrewing the pc to check out the GPU motherboard connection

 

[Aeacus]

but I didn't succeed, could there be any other software or alternative around that can tweak anything in this P3000 by even 10 percent

There aren't many GPU OC programs out there, only handful,
further reading: https://www.digitaltrends.com/computing/best-gpu-overclocking-software/

Do note that they all are created with consumer GPU in mind (Nvidia GeForce, AMD Radeon and maybe even Intel ARC). But none of them may not support workstation GPUs (Nvidia Quadro, AMD Radeon FirePro). So, you can try, but don't get your hopes up.

or could it be that my GPU connect to the motherboard is loosely held or something could be a miss in that connection, I'm even thinking of unscrewing the pc to check out the GPU motherboard connection

Unless you've forcefully rattled your laptop around, i don't see how there could be loose connections.
Laptop internal connectors are well built and resistant to rattling around (since laptops are portable and meant to be carried along).

You can open it up if you like, but do note that you poking in it may make things worse. You could brick the entire laptop, without you being able to power it on anymore. So, there are risks.

I'm just trying anything that can give me some slight hopes

When GPU is faulty (which has been confirmed on several instances), you hoping to high heaven will not make it work again.

Unless you are an Ork from Warhammer 40K, where when you believe something is true, then it will actually become so, brought into being in realspace through the power of the Immaterium by your gestalt psychic ability. Otherwise - No.
Willpower alone doesn't make things work like you want them to.

It is established that when you disable P3000, the system works fine off from HD 630. Why not use that to complete your task?
Sure, HD 630 has FAR slower rendering times compared to P3000, but having 10% of something is better than 100% of nothing.

I'd suggest that you look towards new laptop, rather than trying to resurrect the faulty GPU. Your attempts to get P3000 to work again is like flogging a dead horse. There is no point in doing that. It wastes your time and effort. Easier is getting a new PC.

With new laptop, it is up to you, if:
* buy another used laptop, like you did with this one, when you bought 6.5 year old used laptop and were able to use it 1.5 years, until GPU gave out.
or
* buy brand new laptop and use it for ~8-10 years issue free.

 

[oliveria]

Thanks a lot for your advice, this time I'll try a new laptop that I can use for longer, I also think the reason as to why the ebay seller was selling this old machine that's giving me hell could be because he already saw signs of a failing graphics card and decided to put it on sale

 

[Aeacus]

I also think the reason as to why the ebay seller was selling this old machine that's giving me hell could be because he already saw signs of a failing graphics card and decided to put it on sale

Could be.

Many people put their old and nearly failing hardware on sale. This is most often seen with GPUs, namely ex-mining GPUs, where GPU lifespan is 90%-95% gone and miner puts GPUs to sale, to milk every last bit of money out of them.
I'd say that every RTX 30-series GPU sold as 2nd hand, is ex-mining GPU. Since when RTX 30-series launched, cryptocurrency mining was in full swing.
As of now, cryptocurrency mining with GPUs has fallen off the cliff, after when many crypto brands made so that their coins can't be mined with consumer GPUs. Only with CPUs or dedicated cryptomining GPUs (e.g Nvidia has made such dedicated GPUs for miners).