Question Help with "Kernel 41" error ?

[JLW123178]

I have been experiencing random Kernel 41 category (63) reboots ever since I upgraded from a powercolor radeon 6700 xt to a Reference Radeon 6950 xt. When I upgraded I also bought 2x8 sticks of ram. This has been going on for the past two weeks. The error only happens when I'm gaming during either graphic intense times, loading or even just playing a scene. None of this happened while I had the radeon 6700 xt. Here is my pc setup. I'm beginning to think that the issue is that I was unlucky and got a new GPU bad out of the box. I had a similar experience when I had the radeon 5600 gpu. I've only had the PSU since 04/2022 so almost a year old. Overall I can be watching youtube or internet surf all day with no issues. But while gaming these Kernel 41 errors will randomly happen anytime. When I look for the dump file turns out that none was made. Also I've noticed no spikes in voltage or heat once this error happens. When the pc reboots the AMD adrenaline software does not show the wattman error. Any tips will be appreciated. I'm going back to microcenter to exchange the GPU since I have the extended warranty. Hopefully I can replace it with the same GPU. If not I might go and try a NVIDIA 3000 series.

CPU: AMD RYZEN 5600X
MOBO: MSI X-570 PRO
RAM: 4x8 RIPJAWS G.SKILL 3600
Storage: TWO M.2 1TB DRIVES, 1 TERABYTE SSD
GPU: REFERENCE RADEON 6950 XT
PSU: CORSAIR 80+ PLATNINUM HX1000W
WIN 10

Here are the trouble shooting steps I've done already

DDU removal, in safe mode, and installation of radeon drivers
Removed and reseated the gpu
Disconnected and reconnected gpu power cables
Disconnected and reconnected all power cables to PSU
Removed the two new sticks of ram
Performed memory test and found no error
Nothing overclocked and no xmp profile on
Performed stress test on GPU, CPU, and PSU with no crash and no errors reported
All chipset drivers and BIOS up to date, also reinstalled latest chipsest drivers and BIOS

 

[jojesa]

The Kernel 41 just means that the system rebooted without cleanly shutting down first.
Does the system shuts down when using only 2 RAM modules?

 

[JLW123178]

The Kernel 41 just means that the system rebooted without cleanly shutting down first.
Does the system shuts down when using only 2 RAM modules?

This happens with just the two ram modules that I have currently and also when I put in the two new ones I bought with the 6950 xt. Right now I have the new ones out and I still get same error. All this started happening when I upgraded from the 6700 xt to the 6950 xt. Never had this issue while using the 6700 xt. I would think maybe it's the PSU, but I don't see how the PSU became bad when I changed GPU's.

 

[jojesa]

Matching different RAM could cause that issue.

Try resetting the BIOS to default.
Power off, unplug power cord and press the power button for 20 seconds.
Install just one RAM kit (the old or the new)
Remove the CMOS battery and wait 5 minutes before placing it back.
Reconnect the power cord and power on.
Let the BIOS taking is time to train the memory, which could take up to 5 minutes.
The system could reboot several times during this process.

I would not recommend mixing and matching RAM from different kits.
The RAM that goes into a RAM kit (package) has been tested to work together. There are times that different RAM modules, even if they look similar, and have the same model they could have different IC chips brands.
That is why manufacturers don’t guarantee mixing or adding one RAM modules to another, even when they are the "exact same" model.

 

[Roland Of Gilead]

Hey there,

Yes, kernal 41, typically means there was a loss of power to the system, so it reboots.

It could be for a number of reasons. If you are experienceing this issue since the new GPU, then it's likely the PSU.

The HX is a really good PSU. I'm not sure if that's the issue, but it could be. How old is it?

Did you clear CMOS after bios updates?

AS @jojesa mentioned, mixing and matching DIMMs can result in the issue you're facing too.

As suggested, try one kit at a time. If it doesn't crash then you know that the DIMMs don't mix that well. You can also consider manually setting mem speeds/voltage etc in the bios. This might get them to work a bit better, if that is the issue.

 

[johnbl]

look in c:\windows\minidump and see if there are any *.dmp files

error 41 is logged on the reboot of a machine that did not complete shutdown.
IE power failure cut off, cpu reset, or system bugcheck
a system bugcheck should try to do a diagnostic dump and store the dmp file

 

[JLW123178]

look in c:\windows\minidump and see if there are any *.dmp files

error 41 is logged on the reboot of a machine that did not complete shutdown.
IE power failure cut off, cpu reset, or system bugcheck
a system bugcheck should try to do a diagnostic dump and store the dmp file

No dump file.

 

[JLW123178]

Hey there,

Yes, kernal 41, typically means there was a loss of power to the system, so it reboots.

It could be for a number of reasons. If you are experienceing this issue since the new GPU, then it's likely the PSU.

The HX is a really good PSU. I'm not sure if that's the issue, but it could be. How old is it?

Did you clear CMOS after bios updates?

AS @jojesa mentioned, mixing and matching DIMMs can result in the issue you're facing too.

As suggested, try one kit at a time. If it doesn't crash then you know that the DIMMs don't mix that well. You can also consider manually setting mem speeds/voltage etc in the bios. This might get them to work a bit better, if that is the issue.

I've already removed the two new sticks of ram and I'm still getting this kernel 41 error/reboot. I've had the PSU since 4/2022 so not even a year. I've had PSU's go bad with my past builds so I'm familiar with bad PSU's. I highly doubt my 1000w PSU went bad just because I changed GPU's. The bios is up to date and everything is set to either default or auto. No overclocking or xmp profile. When I had this issue with the Radeon 5600xt I upgraded to the 6700xt and no more kernel error/reboots. That's why I'm thinking bad gpu. Also since the 6950xt is the only new piece of hardware in my tower then it's plausible that's the hardware issue causing this.

 

[johnbl]

No dump file.

bummer, now you have to look for power/heat related reasons for the reboot.
ie something like gpu heats up, pulls too much power via pci bus, motherboard detects the excess power draw and the motherboard protection circuit resets the cpu until the power draw is corrected. the monitor goes black. you might have to reset the power supply to correct the problem, depend on where the signal came from. cheap power supplies fake the signal to the motherboard, these often result in a bugcheck because the CPU restarts before the power is stable and the CPU cache gets corrupted and the CPU calls the bugcheck. Does not happen with a good power supply, you just get a blank screen.

 

[JLW123178]

bummer, now you have to look for power/heat related reasons for the reboot.
ie something like gpu heats up, pulls too much power via pci bus, motherboard detects the excess power draw and the motherboard protection circuit resets the cpu until the power draw is corrected. the monitor goes black. you might have to reset the power supply to correct the problem, depend on where the signal came from. cheap power supplies fake the signal to the motherboard, these often result in a bugcheck because the CPU restarts before the power is stable and the CPU cache gets corrupted and the CPU calls the bugcheck. Does not happen with a good power supply, you just get a blank screen.

While I was gaming I had the adrenaline overlay up. When the kernel error happened no heat build up or any voltage spikes. The screen just went black. Every time I did a trouble shoot step I unplugged and disconnected the PSU so it was reset every time. If you look at my hardware list the PSU isn't cheap and is a good quality. I had this same issue when I had the Radeon 5600xt. Once I upgraded to the 6700xt no more kernel issues. Since I have the extended warranty at microcenter I'm going to replace it. I'm starting to think maybe something on the GPU pc board is bad and is sending a false error causing this kernel reboot.

 

[johnbl]

While I was gaming I had the adrenaline overlay up. When the kernel error happened no heat build up or any voltage spikes. The screen just went black. Every time I did a trouble shoot step I unplugged and disconnected the PSU so it was reset every time. If you look at my hardware list the PSU isn't cheap and is a good quality. I had this same issue when I had the Radeon 5600xt. Once I upgraded to the 6700xt no more kernel issues. Since I have the extended warranty at microcenter I'm going to replace it. I'm starting to think maybe something on the GPU pc board is bad and is sending a false error causing this kernel reboot.

the PSU does have good reviews.
All I would check is the GPU to PSU connection at both ends since it is a modular connection.
the only normal driver problem that could cause this would be a overclocking driver.
my best guess is the motherboard power protections circuits were triggered and the motherboard reset the cpu. With a good power supply you get a black screen. With a cheap power supply the system reboots and the CPU bugchecks on the reboot. (easy to detect with a cheap PSU)

I once had the same thing happen and I took the GPU out and put it in a machine that had disabled motherboard protections circuits. I thought I was being cleaver and during testing the GPU pulled too much power and the PCI/e slot melted and caught fire. burned up the motherboard and GPU.

I would send the GPU back since you can.
software wise, I would make sure there is no overclocking.
bIos, I would make sure the PCI bus is not overclocked, IE running at 100MHz

gpu Power Draw (TDP): 335W
guess if the new one also has the same problem you could underclock to see if your cpu is still reset.
+++++++++++++
edit:
looks like this power supply has a rail selection switch.
is it set to single rail?
single rail should give the most power to GPU,
multi rail might limit the output of the 12v gpu connector.
and force the PCI/e bus to provide the extra power and trip the motherboard power protection circuits.

from docs:
+12V rail selection switch The HX Series power supplies are equipped with a +12V Rail Selection Switch that gives you the choice of single or multiple +12v rails. In the “SINGLE” position, the full output of the power supply’s +12V rail is available to any and all connectors, while in the “MULTIPLE” position, each individual connector has Over-Current Protection so no more than 40A of current can be delivered on any given cable. Deliver the power how you want.

 

[JLW123178]

the PSU does have good reviews.
All I would check is the GPU to PSU connection at both ends since it is a modular connection.
the only normal driver problem that could cause this would be a overclocking driver.
my best guess is the motherboard power protections circuits were triggered and the motherboard reset the cpu. With a good power supply you get a black screen. With a cheap power supply the system reboots and the CPU bugchecks on the reboot. (easy to detect with a cheap PSU)

I once had the same thing happen and I took the GPU out and put it in a machine that had disabled motherboard protections circuits. I thought I was being cleaver and during testing the GPU pulled too much power and the PCI/e slot melted and caught fire. burned up the motherboard and GPU.

I would send the GPU back since you can.
software wise, I would make sure there is no overclocking.
bIos, I would make sure the PCI bus is not overclocked, IE running at 100MHz

gpu Power Draw (TDP): 335W
guess if the new one also has the same problem you could underclock to see if your cpu is still reset.
+++++++++++++
edit:
looks like this power supply has a rail selection switch.
is it set to single rail?
single rail should give the most power to GPU,
multi rail might limit the output of the 12v gpu connector.
and force the PCI/e bus to provide the extra power and trip the motherboard power protection circuits.

from docs:
+12V rail selection switch The HX Series power supplies are equipped with a +12V Rail Selection Switch that gives you the choice of single or multiple +12v rails. In the “SINGLE” position, the full output of the power supply’s +12V rail is available to any and all connectors, while in the “MULTIPLE” position, each individual connector has Over-Current Protection so no more than 40A of current can be delivered on any given cable. Deliver the power how you want.

I just got back from microcenter and was able to replace with another reference radeon 6950 xt. Fingers crossed this time. My PSU has always been on single rail and still is. I did double checked the power cable on both ends and reconnected it to the PSU and to the new 6950. I looked at my BIOS again and everything as always is still set to default and auto. No xmp profile or OC set in BIOS.

 

[JLW123178]

the PSU does have good reviews.
All I would check is the GPU to PSU connection at both ends since it is a modular connection.
the only normal driver problem that could cause this would be a overclocking driver.
my best guess is the motherboard power protections circuits were triggered and the motherboard reset the cpu. With a good power supply you get a black screen. With a cheap power supply the system reboots and the CPU bugchecks on the reboot. (easy to detect with a cheap PSU)

I once had the same thing happen and I took the GPU out and put it in a machine that had disabled motherboard protections circuits. I thought I was being cleaver and during testing the GPU pulled too much power and the PCI/e slot melted and caught fire. burned up the motherboard and GPU.

I would send the GPU back since you can.
software wise, I would make sure there is no overclocking.
bIos, I would make sure the PCI bus is not overclocked, IE running at 100MHz

gpu Power Draw (TDP): 335W
guess if the new one also has the same problem you could underclock to see if your cpu is still reset.
+++++++++++++
edit:
looks like this power supply has a rail selection switch.
is it set to single rail?
single rail should give the most power to GPU,
multi rail might limit the output of the 12v gpu connector.
and force the PCI/e bus to provide the extra power and trip the motherboard power protection circuits.

from docs:
+12V rail selection switch The HX Series power supplies are equipped with a +12V Rail Selection Switch that gives you the choice of single or multiple +12v rails. In the “SINGLE” position, the full output of the power supply’s +12V rail is available to any and all connectors, while in the “MULTIPLE” position, each individual connector has Over-Current Protection so no more than 40A of current can be delivered on any given cable. Deliver the power how you want.

Been back and gaming for about two and half hours and not a single kernel 41 error reboot. Usually I'd get a random kernel 41 error within the first hour of gaming.

 

[johnbl]

Been back and gaming for about two and half hours and not a single kernel 41 error reboot. Usually I'd get a random kernel 41 error within the first hour of gaming.

so replacing the GPU fixed the problem?
(no underclocking)

 

[JLW123178]

So far so good. Every setting in the adrenaline software is set to default for tuning.

 

[JLW123178]

so replacing the GPU fixed the problem?
(no underclocking)

So far so good. Every setting in the adrenaline software is set to default for tuning.

 

[JLW123178]

So far so good. Every setting in the adrenaline software is set to default for tuning.

I can confirm that the kernel error was caused by a bad gpu out of the box. Since replacing the 6950 xt with another one I haven't had a single crash for the past week. Also the two new sticks of ram have been put back in which also confirms that the new ram kit is good to go.