MAG X870E TOMAHAWK WIFI does not work properly when the temperature rises

[zlayapipetk15c602eb]

System Configuration:

• Motherboard: M/B MAG X870E TOMAHAWK WIFI
• CPU: AMD Ryzen 7 9800X3D
• Memory: G.Skill DDR5-6400 65536 MB PC5-51200
• Cooling: be quiet! PURE LOOP 2 360mm (BW019)
• Graphics Card: MSI PCI-Ex GeForce RTX 4060 Ti Gaming X Slim
• Operating System: Windows 11

1. Unstable performance when temperatures exceed approximately 80°C+-, which is visually indicated by the temperature "freezing" on the LED display on the motherboard. This leads to the water cooling system running at full speed continuously, which can only be reset by rebooting the machine. Additionally, some system applications crash, and AMD Ryzen Master stops working.
2. Memory settings were adjusted from 6000 to the base 4800 MHz and tested for 5–6 hours using Windows Memory Diagnostic and MemTest86 with no errors found. However, during one of the MemTest86 tests, the temperature sensor once again froze.
3. MSI Center does not install correctly – on the HARDWARE MONITOR tab, only the GPU data is shown. Moreover, when the temperature "freezes," the application itself crashes, for example:

Faulting application name: CC_Engine_x64.exe, version: 3.0.0.81, time stamp: 0x67a0896c
Faulting module name: CPU_Engine_x64.dll, version: 2.0.0.71, time stamp: 0x6747db01
Exception code: 0xc000041d
Fault offset: 0x00000000000070fc

1. It is possible to play a game without noticing that the temperature has frozen, and only realize it after exiting the game. Sometimes, the game itself may also crash due to this issue.

 

[zotta.vla15dc02ea]

install HWINFO and check your other temps under load (system, MOS, chipset). do you have any other cooling in your case? these temps should be lower under load (compared tot idle), as the other fans in your case should increase their rpm too and bring in fresh air. second, how much paste did you use? is the AIO installed correctly? check that, maybe reinstall it. I suppose this is a fresh windows install and all drivers are installed, right?

 

[zlayapipetk15c602eb]

for example it looks like this - video added, programs that lose information frequently, temperature and so on after that freeze or fall into error. Sometimes it can work for 5 minutes for example - the temperature does not rise above 88 degrees, AIO works normally. Silent Base 802 case with 3 coolers.

 

[CuNi]

I have the exact same issue, but mine freezing even way below the 85°C.
Came to the forum just now that it froze again, and it's sitting at 52°C according to HWINFO.
I already contacted MSI through a Support Form yesterday, and I am awaiting a response.
I wonder if this is something that can be fixed by a BIOS update, or if the whole Board is faulty.

Once it's frozen, I cannot put the PC to sleep anymore as it will be stuck once it turns off the screen and only a reboot or hard reset brings it back to life, just like the temp sensor/display.

 

[Krisk7157302df]

Same MB, CPU, RAM using a AIO cooler with a RTX 4080 Super GPU using BETA BIOS A41

Mine maxes around 70-72C under full CPU load, then idles at around 38-43C. Not seen any crashes happen, since building it 2 months ago.

 

[zotta.vla15dc02ea]

for example it looks like this - video added, programs that lose information frequently, temperature and so on after that freeze or fall into error. Sometimes it can work for 5 minutes for example - the temperature does not rise above 88 degrees, AIO works normally. Silent Base 802 case with 3 coolers.

I notice from your video that when it gets stuck, all cores are loaded, except core 0. That shouldn't happen in a multithreaded test/benchmark. Might be worth trying with a different cpu if you can or take your cpu and install it in a different board, see if it behaves the same way. at 00:29 in your video, core 0 crashes to 0 MHz, as reported by hwinfo active clock window (on the right).

 

[zlayapipetk15c602eb]

I see it too, but at this point it is also visible how all information on CPU cores stops being displayed at all. Most likely this is the behavior of CPU-z when losing access to data, because AMD master did not show this.
unfortunately I do not have the opportunity to use the swap fund, because then I would have localized the problem much faster

 

[zotta.vla15dc02ea]

AMD Master probably froze completely before it could display this behavior. if you cannot swap somehow either one of them, then try to return the cpu as defective first (it should be under warranty).

 

[zlayapipetk15c602eb]

I'm 99% sure that the problem is in the m/b, since I don't see any other problems, I can play 2-3 hours of "heavy" games and everything works fine (except that the AIO processes the latest temperature), there is only a possibility that programs or services that process or manage CPU or m/b parameters may fail, which is also annoying.
The same MSI center crashes due to loss of access to the data channel in the middle of work, so I was forced to simply uninstall it
Unfortunately, it is quite difficult to return a part, because the seller needs to prove with arguments what exactly is problematic, for example, on a replacement fund, and unfortunately I do not have such an opportunity.

 

[CuNi]

The Board influences how the CPU behaves, and with the Board "crashing" I could see this affecting the CPU too.
If I launch HWINFO after I notice that the board stops updating, it takes a long while "Detecting CPU#0".
Right now the Issue already manifested after 2h of use, and it got stuck on "59,5°C".

It is really hard to pinpoint if it's the CPU or Board that causes this issue.
Is there any definitive way to pinpoint what part is at fault here?
Like Zlayapipetk15 said, except for the Temp-Sensor being stuck, the CPU behaves absolutely normal even after the crash.
It's like the telemetry dies completely, but no way to find out which side of it.

 

[zotta.vla15dc02ea]

did you update to the latest bios? if not, do that too. in order to check if core 0 indeed crashes under load you can try testing just that one with OCCT, setting it like in the photo attached (of course, you can check all the cores one by one like this).

 

[CuNi]

@Zotta, I will try that in a second too, but I did find more evidence for it being the board.

Just now, I ran a CPU-Z Stress test, just like Zlayapipetk15 did.
I let HWINFO monitor all Core-Clocks and the Temperatur.
Don't mind the small dip, I restarted the Stress-Test there around where all cores dropped in clock.

As you can see, while under load, the temp on my CPU fluctuates constantly close to 71.
I saw the Active-Clock show Core0 with 0Mhz in the display like in Zlayapipetk15 Video, but as you can see in the graph, it was not registered which lets me to assume it's just a display bug when HWINFO lost the connection to the telemetry.
You can see where the Temps start to fluctuate and start to be 100% constant. That's when the "Crash" occured.
The whole time, CPU-Z was running the stress and the score did not drip like you would expect it when a core would actually crash or clock down to 0Mhz.

So this brings me to believe it is just an issue with the telemetry and I would assume it's the board that is at fault.
I have no deeper understanding about how the telemetry is being collected, if it is by the CPU itself or thought the MB.
Maybe someone with more knowledge can chime in and explain how Temp and Clock Information are gathered.

I will install OCCT in a second and use that to just stress Core-0.
Would OCCT catch a crash and log it? Or is it like CPU-Z where it doesn't make any logs?

Edit:
To add to this, I'm currently on the BIOS 7E59v2A4

 

[zotta.vla15dc02ea]

you can set in the OCCT settings to stop on error or/and critical temp (but it's not the temp giving you headaches). it logs all steps and you have a button to copy to clipboard and paste in a file if you want to save the info. the info on cpu temps is taken from sensors inside the cpu. in the HWINFO sensors status window, if you hover with your mouse on various cpu temps it will tell you the source of that information.

 

[CuNi]

you can set in the OCCT settings to stop on error or/and critical temp (but it's not the temp giving you headaches). it logs all steps and you have a button to copy to clipboard and paste in a file if you want to save the info. the info on cpu temps is taken from sensors inside the cpu. in the HWINFO sensors status window, if you hover with your mouse on various cpu temps it will tell you the source of that information.

Ah I see.
So far it's been running for 27 Minutes in OCCT only on Core-0 and no issues.
With all-core stress in CPU-Z, the issue normally appeared after a few minutes already.
I'll let it run till 30 Minutes and then change OCCT to All-Core and monitor it again.
The Temp that is read by the Motherboards 7-Segment Display and the one that freezes is the Tctl/Tdie, which seems to be inside the CPU.
This whole situation is so weird.

 

[zotta.vla15dc02ea]

Ok, try all core. Yes, that's why I'd rather put the blame on the cpu. Maybe try reseating the cpu and reinstall your cooling? Making sure you don't overtighten. Do you use a thermalright cpu frame or something similar? Its easy to overtighten those.

 

[CuNi]

I tried all cores now and after 11 Minutes the same thing happened.

OCCT does say it didn't detect any errors.
Note though that as it happened, OCCT also lost telemetry access.
Before the issue, it showed the Power Draw, after it happened, it fell down to 0W.

I'm using an Arctic Liquid Freezer 3 420 AIO.
It's just weird that the telemetry is the only thing affected by this issue.
The rest of the CPU works just fine and no application crashes or freezes because of it, which is what confuses me the most.

I already send a ticket to MSI on Friday evening, so we'll see what they'll say about the matter.
Just to be sure, I'll reach out to AMD too and ask them what their Opinion is of this.

 

[zlayapipetk15c602eb]

Unfortunately, my problem turned out to be in the motherboard, which was shown by tests with a different processor. I agreed with the guys from the computer store and brought the computer to them. We replaced the processor with an AMD 8400f and ran tests. On this processor, telemetry also froze. A little differently - for example, on my 9800x3d, the Power Draw dropped to zero, then on the 8400f it also simply Power Draw froze, but the result showed that the problem is 99.9% in the motherboard

 

[basbar1160302ee]

Did anyone find a solution to this problem other than replacing the motherboard by any chance?
I am facing very similar problems with a 9800X3D on a X870E Tomahawk: CPU temperature freezing (also on the motherboard temperature display); CPU power usage etc. not being reported anymore; some programs not working/starting etc.
At the end it also leads to my PC freezing and shutting down.
Most of the time it happened after running CPU intensive things, but it also happened once without having any high load on the CPU.

 

[CuNi]

Did anyone find a solution to this problem other than replacing the motherboard by any chance?
I am facing very similar problems with a 9800X3D on a X870E Tomahawk: CPU temperature freezing (also on the motherboard temperature display); CPU power usage etc. not being reported anymore; some programs not working/starting etc.
At the end it also leads to my PC freezing and shutting down.
Most of the time it happened after running CPU intensive things, but it also happened once without having any high load on the CPU.

I had to do the same thing as zlayapipetk15c602eb.
MSI Support advised me to RMA the Board to the shop I bought it from, and they could also verify that it was the board that was the issue.
I got a replacement board that works without issue right out of the box.

 

[michelelaurell152d02d4]

I have the same problem. What happens if I bought the PC already pre-assembled online? Do I have to send it back to the store? I use this PC for work. Are you sure it's not a BIOS issue?