PCI express errors/crashes kernel erro41

Associate
OP
Joined
8 Mar 2024
Posts
36
Location
liverpool
You could be on to something if the CPU wasn't stable at oob power settings it could have affected signal integrity between it and a dedicated GPU, though saying that and the behaviour you were getting is very similar to what I was getting, during your crashes before did you ever get in game messages like "pcie device removed, this happens if the GPU is disconnected" I had that plus loads of error 41's in event viewer on my old setup.
No buddy no error messages in game. Pc could be just sitting at desktop idle then black screen all fans hit 100% then check event viewer and it stated it was a kernel error. I don’t think I have an issue with my gpu or my memory. I have tried applying the power limits and since then seems to be stable
 
Soldato
Joined
28 Aug 2017
Posts
2,787
Location
United Kingdom
How would power limits affect it as I say appears to be stable now

The cpu is power limited now a oob chip will pull over well 300w until you manually change the power limit in the bios, for the cpu to reach its max turbo clocks requires a tonne of voltage and current, coupled with the cores boosting to the ragged edge could explain the crashing until you limited the chip, 253w is still a fair amount but much better than auto. A ratio of x54 is still very high and wont see any drop in games vs when it was crashing.
 
Associate
Joined
2 Mar 2003
Posts
1,591
Location
UK
My suggestions might seem a little extreme but your welcome to try them.

Remove GPU assuming your CPU has an IGP.
Remove any other PCIe expansion cards.
Remove all NVMe drives.

Stick Linux on a USB and run a live environment and test to see how the system behaves.

Of course ensure you are running no overclocks.

If all stable add only one NVMe and do a clean build of Windows. If all stable then add your GPU and so on. Due to the nature of the fault it will take time to resolve.

If the PC fails in its minimum configuration then your next step is to somehow rule out CPU and mobo.
 
Associate
OP
Joined
8 Mar 2024
Posts
36
Location
liverpool
I’ve literally upgraded the motherboard good few months back I really seem to think it’s due to voltage causing it to be unstable. Do you think il loose a lot of performance by running it at stock wattage of 253?
 
Man of Honour
Joined
22 Jun 2006
Posts
11,656
Do you think il loose a lot of performance by running it at stock wattage of 253?
In the real world? Unlikely, not many use cases benefit from exceeding the power limit. Some numbers for how performance scales with the power limits below *.

Since you own an extreme overclocking motherboard, if you care about every .5% in a benchmark then yes, the loss might be significant.

*
 
Soldato
Joined
28 Aug 2017
Posts
2,787
Location
United Kingdom
I’ve literally upgraded the motherboard good few months back I really seem to think it’s due to voltage causing it to be unstable. Do you think il loose a lot of performance by running it at stock wattage of 253?

i very much doubt it, 253w is still a lot of power than can be given to a cpu, gaming will see next to no drop because you have a 4090.

to put it into context my 7950x has a max power draw of 230w that it can use (170w tdp is default), but in day to day gaming load i barley see 150w max, most of the time its lower than that, only if i hit the cpu with a heavy load did i see close to 220w but then temps became the major factor, 95c is too toasty for me.

stock wattage for the 14900k is 125w not 253, the bigger number is the wattage the chip can pull under boost conditions, if you lock to 125w then your cpu will run around 3.4ghz which is its base clock.
 
Last edited:
Associate
OP
Joined
8 Mar 2024
Posts
36
Location
liverpool
Can’t really see it being the board as this is the second board I’ve had and as regards to the cpu I’ve posted to Intel support I’ve had nothing but headaches with this pc wish I stayed on console haha
 
Man of Honour
Joined
22 Jun 2006
Posts
11,656
Red advised to alter the svid to intel tailgate what does that actually do?
Do you mean Intel failsafe? I don't know what tailgate would mean. As far as I know, SVID uses a table of frequencies and voltages and the CPU requests from the motherboard VRM what voltage it wants to receive at a given frequency. If you use Intel failsafe, I assume it just uses Intel defaults.

just run a avx2 test using Intel xtu and same issue once again im really stumped now as to what it could be
It is not uncommon to have overclocks that are stable without AVX and unstable with AVX.
 
Associate
OP
Joined
8 Mar 2024
Posts
36
Location
liverpool
Do you mean Intel failsafe? I don't know what tailgate would mean. As far as I know, SVID uses a table of frequencies and voltages and the CPU requests from the motherboard VRM what voltage it wants to receive at a given frequency. If you use Intel failsafe, I assume it just uses Intel defaults.


It is not uncommon to have overclocks that are stable without AVX and unstable with AVX.
Yer it doesn’t make sense to me as been fine for weeks and now just got this issue and as regards to running intels svid failsafe that puts 1.5vcore through the cpu which I don’t really want to be doing as it’s ran fine for weeks on 1.4 plus isn’t 1.5 a bit on the high side and will degrade the cpu faster?
 
Man of Honour
Joined
22 Jun 2006
Posts
11,656
running intels svid failsafe that puts 1.5vcore through the cpu which I don’t really want to be doing as it’s ran fine for weeks on 1.4 plus isn’t 1.5 a bit on the high side and will degrade the cpu faster?
I would assume that a failsafe setting would use higher voltages, because it needs to account for poorer quality CPU silicon and less stable/capable power delivery from the motherboard's VRM.

In terms of degradation, I think it depends on how loaded the CPU is, since if the CPU uses high volts at low loads to achieve high single-core boost clocks, that is less likely to degrade the CPU than using high volts at high load, across all cores.

It is normal for modern high performance CPUs to use high stock voltages to achieve their headline boost clocks, but if the voltage does not fall when you're running an all-core load, that's bad.
 
Man of Honour
Joined
22 Jun 2006
Posts
11,656
Does it sound like a faulty cpu to you?
There's a lot of variables with these high-end CPUs (i.e. quality of the motherboard VRM/power delivery and the PSU, the PSU's reaction to instantaneous load, motherboard settings, cooling and cooling profile e.g. how quickly it reacts to load, the memory and memory settings, the individual workload and how demanding it is), so it is really hard to say.

What wookiee87 said might be right here, that Intel just can't test these CPUs thoroughly enough, or that their test conditions for those boost clocks are too different to their standard usage in the "real world".

It was a lot easier when CPUs only had 4 cores and a small variance in power/heat output, but they can use anything from 30 - 400 watts nowadays and have massive variance in volts and frequencies. Heck, the 14900K effectively has 2 different CPUs/architectures in one. They also have several different boost technologies, the standard turbo boost and TVB and the instability could be anywhere in there.
 
Last edited:
Man of Honour
Joined
22 Jun 2006
Posts
11,656
Yer I agree this is the most temperamental pc I’ve ever built I’m going to see what Intel says because it’s not normal
I think motherboard manufacturers not obeying Intel's default settings "out of the box" (i.e. effectively overclocking everything without anyone knowing) is also part of the problem with these CPUs, but if that has anything to do with your case, I don't know.

 
Back
Top Bottom