CPU damaged?

[Archerion]

So I have bought a new "Ryzen 7 2700x" CPU and a "MSI x470 Gaming Pro" Mainboard. I installed all the components and Windows 10 Home, did the updates and started to install my first games. 

 

After 6h of runtime my PC suddenly went black and basically turned off. I thought "That was weird." and restarted the system. Everything booted and ran for a little bit like before, but I checked on the temps and In idle mode, a tool called "Speccy" told me my CPU was at 100 °C. Now I just thought "WTF?!" and Installed the tools from MSI that came with the Mainboard and after a quick install I checked on the temps there and they where at around about 30 to 40 °C while installing a game. That was much more likely and the CPU cooler, an ARCTIC Freezer XTREME, felt cooler as well. But I checked the Core Voltage and they were at 1.458V at this moment. I dont know if that is normal, but I think that is a little bit high. After that the system started to get unstable or something. When I started to load a program my whole PC freeze for 10 seconds and sometimes got completely stuck, so I have to shut down the pc and restarted. Now it sometimes restarts normally, but then has these freezes or it gets stuck while loading Windows.

 

Might be that my CPU is a underperforming one and got damaged because of that or the Mainboard is bad, idk. So now it is shut down for safety.

I don't necessarily want to overclock and now I think I should "underclock" it. Should I turn it down to 1.2 Volts and the frequency accordingly, or should I send the CPU and/or the mainboard back befor I do more damage to it?

[Semper]

1.458v is a stock voltage profile so far as I know for the 2700x. It's the higher end of what AMD has deemed the silicon is capable.

A direct quote from AMD on the matter:
 

Quote

Robert Hallock from AMD explained this perfectly.

 

The CPU is programmed to use these voltages automatically. We know it's safe, because we designed it that way. The CPU cannot and does not use voltages that are unsafe for the silicon.

The key thing that people forget in these cases is time and temperature. Running 1.4V or 1.5V here and there is not a big deal, because the CPU will eventually back down according to its pre-programmed model. Or if you have great cooling, that also offsets the thermal effect of voltage. In either case, you're seeing momentary blips of voltage that are offset by the hours per day your CPU is probably doing nothing at all--at a very low voltage.

The average vcore for Ryzen over time is around 1.25V (give or take).

tl;dr: leave the CPU alone, let it do its thing, don't worry. We designed the CPU to do this.

I'm getting conflicting reports on thermal throttling and thermal shutdown for the 2700x, but it would appear that it either throttles or shuts down at 95c, or potentially throttles at 95, and shuts down at 110. I don't own this chip so i'm unable to confirm either way. I believe this to be the cause of your black screen.

I would recommend that you check your heatsink and/or thermal paste for fitting issues. it seems like it's not adequately cooling the chip.

[Automobili3XF]

31 minutes ago, Archerion said:

Speccy

Try HWiNFO or Ryzen Master.

[Archerion]

I checked the cooler for tightness and it seams to be alright and the cooler came with paste, so I used this one, so It should be spread nice and evenly. Like I said, I used the mainboards tools for checking the temperature as well, which were at 40°C. I could try HWiNFO later. The thing is that it ran for 6h without any problems and then they appeared, like it was getting to much Voltage, because average seems to be at 1.45 with mine. I have the stock cooler as well, so I could try this one, after I checked with HWiNFO. If that doesn't help, should I then lower the voltage and the frequency?

[Sovek]

you sure its not a software problem? I've recently wiped and reinstalled windows and now I've had two crashes, freezes and restarts. This is with a Ryzen system

[Wolly9102]

11 hours ago, Archerion said:

I checked the cooler for tightness and it seams to be alright and the cooler came with paste, so I used this one, so It should be spread nice and evenly. Like I said, I used the mainboards tools for checking the temperature as well, which were at 40°C. I could try HWiNFO later. The thing is that it ran for 6h without any problems and then they appeared, like it was getting to much Voltage, because average seems to be at 1.45 with mine. I have the stock cooler as well, so I could try this one, after I checked with HWiNFO. If that doesn't help, should I then lower the voltage and the frequency?

 

1,45 is fine. Mine does the same thing. I've noticed it get's to 1,45 by single thearded workloads. If run a stress test on all cores it's runs itself at 1.3. I've added some screenshots while the CPU is doing different workloads

 

 

[Wolly9102]

[Wolly9102]

12 hours ago, Semper said:

I'm getting conflicting reports on thermal throttling and thermal shutdown for the 2700x, but it would appear that it either throttles or shuts down at 95c, or potentially throttles at 95, and shuts down at 110. I don't own this chip so i'm unable to confirm either way.

You were almost correct  It throttles 85 and shuts down at 115

[Archerion]

Could be a software problem. I checked the... I think it's error log in englisch... in Windows and there was something with Kernel error. So that could be something.

I will search for some solutions. Cooling is fine HWiNFO showed  40°C so everithing is good with the cooler.

[Archerion]

The kernel error was mostly 41, which just means, that the System wasn't shut down correctly. For other errors I have to look, when I'm back home again.

[Archerion]

Ok Guys n' Girls, it most likely isn't a software problem... I installed Win 10 completely new.... ran for 5 minutes... and blackscreen. It ran further and then it showed me a small white box with my mouse cursor in it. Now it won't start up anymore... I mean it will try to load Windows but 5 secs into the loading cycle it freezes and that's it.

 

I can Install windows from my usb drive onto the ssd and there is no issue at all, but when it's loaded, well it dies in some minutes.

The only things I can check is the graphics card, the pcie riser, the power supply and the hdd with my old system. I hope I can find the error in the pcie raiser cable.

[Archerion]

Update:

I have  gone through testing many things, and I'm now starting to think that the actual problem is the SSD. I have an 250GB Samsung EVO 970 M.2 2280. I will do an CPU stress test today, to verify that everything is ok there and I will try to install Windows onto the SSD again, and see if it works then or if it fails.

 

I'd like to test the SSD, with an test tool as well, but I don't know if that one is siutable: https://www.smartmontools.org/

Do you have any other test tools for me that work good? The problem is, that it needs to be bootable, because the OS isn't loading.

 

By the way, do you have an Idea why an SSD breaks that quickly? It should just have run Windows at this moment, while I was playing a game and was installing one which are all on the HDD.

[Archerion]

Problem pre-solved:

In short: The problem was the PCIe-Riser, that I used to separate the GPU and its heat from the CPU.Or at least I think that it was the problem, because now it seems to run stably and loads correctly. I just let it run for an hour yesterday evening, but I think it's fine now.

 

I just didn't come to the solution for so long, because the GPU sent signals just fine when I was in the UEFI or loading some bootable test programs. It seems that the PCIe-Riser just didn't sent any Signals to the GPU, when it loaded Windows... which I don't really get... but ok. As an last option I just plugged the GPU in without the Riser, because it was the last thing, that I didn't check.But like many times Its always the last option, like In my childhood, when I played with my Nintendo DS... the game that I wanted to play was often the last game, that I pulled out of the pocket.

 

My last question is: How can it be that an PCIe-Riser stops working after 6 hours, without anything changing to it?

 

Anyway, if it still should be something I'll let you know.