Event 55, Kernel-Processor-Power Errors?

S

supra21

n00b
Joined
Aug 13, 2021
Messages
38
| 4090 FE | Corsair H100i | EVGA T2 1000w |
| Samsung 990 Pro 2TB | Aorus GEN4 2TB | Corsair MP510 1TB |
| Aorus x570 Master | Ryzen 5800X3D | TG 8Pack Edition 32GB |

I've not had any issues with my pc, but as of last week have been getting this Event 55 error when gaming (cyberpunk 2077), can be after 30 mins or sometimes after 3 hours but pc will just power off. I power it back up and it takes a few mins and bios reverts to default settings. I've tired again with just standard PBO and XMP enabled, still crash. My 4090 was curve optimised with afterburner and I had a +1200mhz OC on memory with power limit of 80% (350w). Both CPU and GPU temps never exceeded 65 degrees when playing CP2077.
I reduced mem OC to +1000mhz still crash. I think I also tried default GPU settings.
Updated AMD chipset drivers to latest version, nvidia drivers latest version, windows updated (Win 11 23H2), Bios updated to latest version (F37h).
Ran the sfc/scannow and following DISM commands all comes back clean no errors.
Ran an OCCT stress test on CPU for around 30 mins no crash.

General desktop usage browsing/videos have not encountered a crash. Googling seems to shoot me off in many different directions so just need some opinions on a method of finding the cause of the issue. The only thing that comes to mind is when I was switching out my old CPU for the 5800x3d I did get a tiny bit of thermal paste in the socket (this was around a year ago). Other than that is my PSU potentially at fault here?

I get about x16 of these logs for each "processor":
Processor 0 in group 0 exposes the following power management capabilities:

Idle state type: ACPI Idle (C) States (2 state(s))

Performance state type: ACPI Collaborative Processor Performance Control
Nominal Frequency (MHz): 3401
Maximum performance percentage: 150
Minimum performance percentage: 51
Minimum throttle percentage: 16
Click to expand...

Reviewing the Logs again also getting Event ID 125:
ACPI thermal zone \_TZ.PCT0 has been enumerated.
_PSV = 290K
_TC1 = 0
_TC2 = 0
_TSP = 1000ms
_AC0 = 0K
_AC1 = 0K
_AC2 = 0K
_AC3 = 0K
_AC4 = 0K
_AC5 = 0K
_AC6 = 0K
_AC7 = 0K
_AC8 = 0K
_AC9 = 0K
_CRT = 294K
_HOT = 293K
minimum throttle = 0
_CR3 = 0K
Click to expand...

Any suggestions would be appreciated.
 
What case you got that in? fan config? Seems like whatever is in that thermal zone is getting hot (HOT, CRT), the system may be shutting down to save that component.
 
Nobu said:
What case you got that in? fan config? Seems like whatever is in that thermal zone is getting hot (HOT, CRT), the system may be shutting down to save that component.
Click to expand...
Meshify C
x2 corsair LL 140mm front intake
x1 120mm noctua bottom intake
x2 120mm corsair LL top exhaust (aio rad)
x1 120mm corsair LL rear exhaust

Have all fans connected to corsair commander pro with icue controlling fan speeds.
 
That seems fine. Probably not the paste, I would think it'd have issues much more frequently or immediately would fail.

Could be power. You do have three large SSDs, a ton of ram, and a powerful video card you have overclocked. A short spike in demand while gaming could trip ocp on the psu and cause it to hard shut off.

Check all the power cables inside your system, and ensure the AC cable is in good shape and fully plugged as well.

Oh, some users reported issues with iCue causing crashes and shut-downs, so might be worth disabling it.
 
Last edited:
I just ran a few benchmarks in succession, then upped the render res to 5431x3055 that bench went through fine. Then upped it to 7680x4320 and powered off just as the bench was about to start so it does seem to point to OCP tripping, but this still occurs even when I've power limited my card to 350w?
I was logging with HWinfo in the background: LOG

I will double check the cables in the case, AC cable is good and fully plugged in directly into wall socket.

Ah the dreaded iCue, I am running and older version (4.33.138) due to v5 just not working well with my older corsair hardware. Looking over change logs seems to have been many changes so perhaps i'll update to latest version see if that makes any difference if not then uninstall it and re-test.
 
I thought updating icue solved it got through 7 runs of the benchmark at 8k, then on the 8th it powered off.
Edit: after uninstalling icue ran multiple cp2077 benchmarks in succession at different resolutions, temps were warmer as fans were just at low speed, but no random shutdowns...
Before I get too excited will need to have a few play sessions to verify if iCue was the cause but it looks positive for now.
 
Last edited:
Well completely disconnected any icue controllers, plugged all fans into mobo headers and still randomly powers off while playing CP2077. Today I disabled XMP left PBO and GPU OC enabled have just been playing CP2077 for 3 hours with no issue. I have the Team Group 8Pack Edition 32GB (2x16GB) TDPPD432G3200HC14ADC01 Kit and have been running them at these settings for quite a while prior to issue occurring: https://hardforum.com/threads/bdie-...am-oc-timings-advice.2027537/#post-1045644957
Thing is I am still not sure if this is the issue either as I am sure I had an instance where the power went off with cpu and ram with mobo defaults while playing CP2077. I tired running Kombustor stress test at 8k with OC enabled and powerlimit at 90% (400w+) for around 15 mins and no power off. I cannot seem to find a guaranteed way to trigger the issue making this somewhat of a nightmare.
 
Last edited:
first test the cmos battery, its about that time. second, set everything to default, no ocs. then third, look at the psu. 1000w is enough if its new enough, if not the ocp might be tripping on power spikes. happened to kyle...
 
I will leave my pc unplugged tonight to check cmos battery. I've set everything to default. The PSU was purchased in 2019, so not that old. Also a question can a spike still occur if gpu is power limited via afterburner? as in my card has been limited to 350w up until this point, so can the 4090 spike enough to trip OCP?
 
I've had this exact same problem now a couple of times on my PC and once on someone else's. In each case it was due to the power supply no longer being able to handle the sudden power draw of the GPU. I'd almost bet money that if you run a GPU-Z log and cause the crash you'll find that in each case the system will crash at precisely the same wattage each time. In the log you may see power numbers after the peak draw but it will fall off a cliff after that before the logging terminates.

In my case, it was around 347w. Cyberpunk 2077 was literally the only game that ever caused the GPU to draw that much power and do so that quickly. It's the spikes in draw that cause the issue. Some PSU's can't handle that long term. I have had this issue with the Corsair AX1200 and ASUS ROG Thor 1200w. I can run any other game on the same hardware with those and they'll be fine. The 4090 is better about this than the RTX 3090's are, but it wouldn't surprise me if it still happens, especially in high end configurations and when overclocking is introduced into the mix.

Just my theory, but I've now seen this twice with my own rig and once with someone else's.
 
supra21 said:
Apparently I don't have permission to read that post, lol.
Click to expand...
time to upgrade!
1708863464147.png

"What are the odds of having two 1000W power supplies, from different makers, that exhibit the exact same failure on a 600W load?
I would not have taken that bet.
So, $1100 later, the very first component I replaced, was the point of failure all along."

kyle had 2 1200w psus, one NIB, tripping ocp on the power spikes from a 6900xt. totally possible in your situation too.

how many power cables are you running to the gpu?
 
Well during a 30 min play session in CP2077 today with everything set to default, no OC it powered off...
I am running a cablemod 12v-hpwr to x4 8-pin cable, this one: https://store.cablemod.com/product/...wr-pci-e-cable-for-evga-g7-g6-g5-g3-g2-p2-t2/

Although default settings the 4090 was pulling more wattage (400w+) than with my OC settings which was limited to 350w. With my OC profile I did run the curve optimiser so voltage was higher to hit the higher core clocks in spite of the lower power limit? What I want to know is the level of spike inline with the power limit? (so is a spike with a 4090 OC'd but power limited to 300w the same as a spike with a 4090 not power limited/default).

Either way it's looking to be my PSU. Which is frustrating as I did intended to do a new build this year or next and am awaiting ATX3.1 PSU's to come to market.
 
I got a day off work and was getting ready to potentially pull the trigger on a MSI MEG Ai1300P. But thought let me see if I can find the wattage at where the trip occurs as suggested by Dan_D.
So I started running benches at 5k then 8k (DLSS balanced/perf) but noticed vram kept getting filled up way to quick affecting consistent power draw during bench runs. I have been running the "HD Reworked Mod Ultra Quality" mod which is quite heavy on vram, fine at 4K but above that no go (or so I thought). I disabled that mod and used 4K DLAA for maximum load without filling up vram. Ran the benchmark in succession for 30 mins straight, pulling at peak 434w and no shutdown / OCP trip. Here is the: GPU-Z Log

I dunno if this is just pure chance or the random spikes only occur in game where load is more varied but prior to this I have been able to trigger Power Off / OCP trip running the benchmark in succession with power limited to 350w. So I am thinking the mod has something to do with this, but I am not sure how it would cause this? A CTD from a mod I would expect not an OCP trip. That being said I've been running this mod since last year but the game has also been updated twice since then as well.

To be sure will need to play the game for 100% verification.
 
Last edited:
It's possible a crash caused by not enough vram could cause a sudden drop in power usage, which would cause a sudden rise in voltage on the 12v rail. If it goes up too much, that could cause over voltage protection to trip, although I don't know if the PSU would shut off like for OCP or for a short -- I guess it depends on how OVP is implemented in the power supply.

Very much conjecture, would need to test with a dmm or oscope to be sure.
 
supra21 said:
I got a day off work and was getting ready to potentially pull the trigger on a MSI MEG Ai1300P. But thought let me see if I can find the wattage at where the trip occurs as suggested by Dan_D.
So I started running benches at 5k then 8k (DLSS balanced/perf) but noticed vram kept getting filled up way to quick affecting consistent power draw during bench runs. I have been running the "HD Reworked Mod Ultra Quality" mod which is quite heavy on vram, fine at 4K but above that no go (or so I thought). I disabled that mod and used 4K DLAA for maximum load without filling up vram. Ran the benchmark in succession for 30 mins straight, pulling at peak 434w and no shutdown / OCP trip. Here is the: GPU-Z Log

I dunno if this is just pure chance or the random spikes only occur in game where load is more varied but prior to this I have been able to trigger Power Off / OCP trip running the benchmark in succession with power limited to 350w. So I am thinking the mod has something to do with this, but I am not sure how it would cause this? A CTD from a mod I would expect not an OCP trip. That being said I've been running this mod since last year but the game has also been updated twice since then as well.

To be sure will need to play the game for 100% verification.
Click to expand...
In my case, the shutdowns wouldn't usually occur until I'd been playing awhile. I don't know if the unit had to get hot enough or what was going on there but it didn't do it every time at first and then the intervals between crashes got shorter and shorter. The units also worked fine in machines with less powerful video cards, etc. Capacitor aging, etc. may very well have contributed to it. One unit was 5 years old and the other was about 2 before it started to have problems with that game.
 
Just stopped playing after around 50mins of normal gameplay driving around ncpd missions different areas of the map, one story mission and no OCP yet. Too tired to play any longer so will update this thread once I've had more 2hr+ sessions playing the game. I will say at default settings having a look at the log with this session didn't see the wattage exceed 392w.
 
Downloaded the new patch played for a while, decided to test Path Tracing to see if it had improved after a few mins powered off (I only play with standard RT usually), gpu-z log showing 322w when it went off. The only difference is this time I had a 70% power limit in place (as the perf is the same) which makes no sense to me. Reverted back to 100% power limit and normal RT, been playing for 2 hours no issue and pulling around 380w avg.
So i'm still at a loss as to figuring out what is causing this. Reducing power limit but increasing load (power limit to 70% + Path Tracing) caused machine to power of. Using default power settings + Standard RT with much higher wattage seemingly no issue. I guess the next test is to play with 100% power limit + Path Tracing to see if that triggers the issue?

Just for good measure, I've unplugged and reseated all vga 8-pin cables at psu end, reseated 16-pin in vga, and reseated the ram.
 
Last edited:
supra21 said:
Downloaded the new patch played for a while, decided to test Path Tracing to see if it had improved after a few mins powered off (I only play with standard RT usually), gpu-z log showing 322w when it went off. The only difference is this time I had a 70% power limit in place (as the perf is the same) which makes no sense to me. Reverted back to 100% power limit and normal RT, been playing for 2 hours no issue and pulling around 380w avg.
So i'm still at a loss as to figuring out what is causing this. Reducing power limit but increasing load (power limit to 70% + Path Tracing) caused machine to power of. Using default power settings + Standard RT with much higher wattage seemingly no issue. I guess the next test is to play with 100% power limit + Path Tracing to see if that triggers the issue?

Just for good measure, I've unplugged and reseated all vga 8-pin cables at psu end, reseated 16-pin in vga, and reseated the ram.
Click to expand...

In my situation it wasn't about total power consumption but rather a sudden spike in power draw from the GPU that caused the power supply to shut off. 4080's and 4090's aren't as bad as 3090's are in this regard but that's what your issue seems like to me. A lot of power supplies can't handle that long term.
 
Dan_D said:
In my situation it wasn't about total power consumption but rather a sudden spike in power draw from the GPU that caused the power supply to shut off. 4080's and 4090's aren't as bad as 3090's are in this regard but that's what your issue seems like to me. A lot of power supplies can't handle that long term.
Click to expand...
I assume these spikes did not appear in the log?, because i'm not seeing anything above around 400w.
Just was playing for around 20 mins, Path tracing enabled, 100% power limit and powered off.

Ok, reading this it makes sense, https://www.tweaktown.com/news/8571...ower-excursions-use-up-to-1800w-of/index.html
 
Last edited:
Just bit the bullet and purchased a MSI MEG Ai1300P, should be arriving this evening...
 
supra21 said:
Well during a 30 min play session in CP2077 today with everything set to default, no OC it powered off...
I am running a cablemod 12v-hpwr to x4 8-pin cable, this one: https://store.cablemod.com/product/...wr-pci-e-cable-for-evga-g7-g6-g5-g3-g2-p2-t2/

Although default settings the 4090 was pulling more wattage (400w+) than with my OC settings which was limited to 350w. With my OC profile I did run the curve optimiser so voltage was higher to hit the higher core clocks in spite of the lower power limit? What I want to know is the level of spike inline with the power limit? (so is a spike with a 4090 OC'd but power limited to 300w the same as a spike with a 4090 not power limited/default).

Either way it's looking to be my PSU. Which is frustrating as I did intended to do a new build this year or next and am awaiting ATX3.1 PSU's to come to market.
Click to expand...
If the PSU swap did not work I would swap the Cablemod set for the stock PSU cables, I had intermittent error 41 code reboots that were due to my Cablemod connector pins not making good contact within the connectors. If I touched the wiring harness it would reboot.

https://cdn.wccftech.com/wp-content/uploads/2023/07/youtube-video-gif.gif
 
Last edited:
drutman said:
If the PSU swap did not work I would swap the Cablemod set for the stock PSU cables, I had intermittent error 41 code reboots that were due to my Cablemod connector pins not making good contact within the connectors. If I touched the wiring harness it would reboot.

https://cdn.wccftech.com/wp-content/uploads/2023/07/youtube-video-gif.gif
Click to expand...
The new PSU has a native 12vhpwr port so using the included 12vhpwr cable that came with the new psu. I cba to test again at this point. Additionally my cablemod cable was nowhere near as loose as that adaptor in that gif you linked, also had the glass side panel pushing up against it slightly so no chance of it coming loose over time. I did reseat the cable both at the psu and gpu end during the troubleshooting.
Although I do wish I tried using the the included 12vhpwr adaptor that came with my gpu before buying a new psu.....
Perhaps I could try the cablemod cable in the new PSU at some point.
 
supra21 said:
The new PSU has a native 12vhpwr port so using the included 12vhpwr cable that came with the new psu. I cba to test again at this point. Additionally my cablemod cable was nowhere near as loose as that adaptor in that gif you linked, also had the glass side panel pushing up against it slightly so no chance of it coming loose over time. I did reseat the cable both at the psu and gpu end during the troubleshooting.
Although I do wish I tried using the the included 12vhpwr adaptor that came with my gpu before buying a new psu.....
Perhaps I could try the cablemod cable in the new PSU at some point.
Click to expand...
I'd leave it be, mark the cablemod cable as suspect in case you're tempted to use it in the future. Those connectors are fragile, I would disconnect them as few times as possible.
 
Just reporting back that since the PSU switch over have not had any issues :). Matter of fact I think my system as a whole is more stable, not experiencing the sometimes rare random issues I used to.
Thanks again for all the suggestions. Gonna see if I can RMA my evga PSU and sell it later.
 
Had a 41 kernel and constant reboot error and after HW swaps it was my cablemod PSU cable that was faulty.
 
drutman said:
Had a 41 kernel and constant reboot error and after HW swaps it was my cablemod PSU cable that was faulty.
Click to expand...
When doing my new build I will be testing to verify this. Right now cant be asked to take pc apart again.
 
To clarify it was the PSU to GPU PCI cables that were bad, the rest of the Cablemods were fine. I tried every HW swap and SW updates/uninstalls with no remedy. I used the cables that came with the PSU and has been trouble free for over a year.
 
You must log in or register to reply here.
Back
Top