Computer shuts off automatically under full CPU load

[chiafarming]

Hi, I recently built my first computer and am stuck. I built this computer to farm Chia, a cryptocurrency. The farming process uses the CPU, RAM, and read/write to an SSD.

The first phase of farming is especially CPU intensive. About 5 minutes in, when the CPU is being fully utilized, my computer just crashes and restarts. I've tried this a dozen times over the last few weeks making various changes and this always happens.

What I've checked so far:

• My first thought was that this had to do with the PSU. I tried a brand new different PSU and the same issue happened.
• Second thought was that my CPU cooler wasn't good enough and that the CPU was overheating. So I upgraded to a liquid CPU cooler AOI kit, and the same issue happened.
• Third thought was playing with overclock settings in Bios, but it freezes/shuts off every time I try to modify those settings

Here are the build components:

• Intel Core i9-11900K Rocket Lake 8-Core 3.5 GHz LGA 1200 125
• MSI Z590-A PRO LGA 1200 Intel Z590 SATA 6Gb/s ATX Intel Motherboard
• Team T-FORCE VULCAN Z 64GB (2 x 32GB) 288-Pin DDR4 SDRAM
• Seasonic FOCUS GX-550, 550W 80+ Gold, Full-Modular
• NZXT Kraken X63 RGB RL-KRX63-R1 280mm All-in-One Water Cooler
• Sabrent 4TB Rocket 4 Plus NVMe 4.0 Gen4 PCIe M.2 Internal SSD
• NZXT H510i - Compact ATX Mid-Tower PC Case

I'm running Ubuntu Desktop 18.04 on the computer.

Any help with this is greatly appreciated!

Thanks for your time.

 

[RJARRRPCGP]

So Chia is like X264'ing. (where X264 is CPU-heavy with second-pass video encoding)

Now, you should try to see if the kernel panic is because of a machine check exception. Such symptoms are usually symptoms of an unstable CPU core OC.

Rare to happen at stock and that usually means RMA time!

 

[DeathtoGnomes]

whats the linux equivalent to windows event viewer, thats where I'd start, with error messages.

 

[RJARRRPCGP]

whats the linux equivalent to windows event viewer, that where I'd start, with error messages.

I would check the log files in /var . Look for what looks like a log file named "kernel" or similar.

 

[ThrashZone]

Hi,
I'd guess the mother board is bad seeing it's dying in bios.
Try one stick of memory in the proper single channel slot installed if it crashes try the other stick....

 

[Bill_Bright]

I agree with ThrashZone - try one stick of RAM at a time.

No graphics card? If not, then 550W is plenty big for the PSU - especially with a quality PSU like the Seasonic. But if you are running a graphics card and forgot to list it above, depending on the card, 550W may be underpowered.

 

[chiafarming]

whats the linux equivalent to windows event viewer, thats where I'd start, with error messages.

I would check the log files in /var . Look for what looks like a log file named "kernel" or similar.

Thanks a lot for your replies! I just did a fresh boot run, launched the Chia farming, and the same crash happened. This time I saved the entire log file, I wasn't sure how to narrow it down to just Kernel: https://pastebin.com/AMhq7cWb

Hi,
I'd guess the mother board is bad seeing it's dying in bios.
Try one stick of memory in the proper single channel slot installed if it crashes try the other stick....

Thanks for the help, trying this now & will report back

I agree with ThrashZone - try one stick of RAM at a time.

No graphics card? If not, then 550W is plenty big for the PSU - especially with a quality PSU like the Seasonic. But if you are running a graphics card and forgot to list it above, depending on the card, 550W may be underpowered.

Yep, no graphics card, so PSU definitely shouldn't be the issue

 

[Bill_Bright]

Yep, no graphics card, so PSU definitely shouldn't be the issue

Since you already tried another, I agree.

 

[chiafarming]

Thanks for the help, trying this now & will report back

Update on the single slot RAM: still getting the same crash unfortunately

 

[jboydgolfer]

PSU
Over Volt Protection?

i had to replace my Seasonic Focus+ Gold 750W due to the same issue, except it was high GPU Wattage draw
full system crash & reboot, no error warning, or messages on screen, just like the power went out in an outage just for a second or two.

 

[Bill_Bright]

except it was high GPU Wattage draw

Ummm, what do you mean by this? Components draw current (not wattage, BTW) and only what they need. They could not draw too much unless the graphics solution was faulty (had a shorted component in its circuitry) or the PSU was underpowered (the wrong size) from day 1, or the PSU was faulty and could not deliver what it was supposed to. And if faulty, it would be to all the component (the total current draw) using that voltage, not just the GPU. This is because the power delivered to the graphics solution from the PSU is NOT on a discrete rail. It may have an additional cable connection, but again, that is not a discrete or separate rail.

It is important to note components only pull from the power supply what they need, not what the PSU is capable of providing. That is, if the computer (CPU, motherboard, RAM, drives, fans and graphics card) require 400W, they will pull from the power supplies 400W, regardless if the supply is a 500W supply or a 1000W supply. And in the case of a "Gold" supply @90% efficiency, they will pull from the wall ~445W (445 x .9 = 400.5W) where that extra ~45W is wasted in the form of heat.

 

[RJARRRPCGP]

just like the power went out in an outage just for a second or two.

Possibly a bad capacitor. (a.k.a. cap) It could be a bad APFC filter capacitor! (active power factor correction) (possibly in the power factor correction circuit)

Thanks a lot for your replies! I just did a fresh boot run, launched the Chia farming, and the same crash happened. This time I saved the entire log file, I wasn't sure how to narrow it down to just Kernel: https://pastebin.com/AMhq7cWb


Thanks for the help, trying this now & will report back

View attachment 201386


Yep, no graphics card, so PSU definitely shouldn't be the issue

I can't find a kernel panic, looks like it's elsewhere in /var.

 

[Nike_486DX]

Hey there, did u make sure to have a dedicaded ssd for the os? Cuz chia should run as independent from the os as possible, otherwise it wont work reliably. Also there should be an hdd (ideally 1 ssd + 1 hdd for each plot). So that might be the issue in ur case.

Apart from that, is ur cpu running at stock clocks and voltage? Double check that in bios (start with full cmos reset), also disable everything oc-related (except auto xmp). U may also disable Turbo Boost just to make sure.
Also, check the ssd usage & temps when the issue happens, cuz urs doesnt even have a heatsink intsalled.

 

[chiafarming]

Hey there, did u make sure to have a dedicaded ssd for the os? Cuz chia should run as independent from the os as possible, otherwise it wont work reliably. Also there should be an hdd (ideally 1 ssd + 1 hdd for each plot). So that might be the issue in ur case.

Apart from that, is ur cpu running at stock clocks and voltage? Double check that in bios (start with full cmos reset), also disable everything oc-related (except auto xmp). U may also disable Turbo Boost just to make sure.
Also, check the ssd usage & temps when the issue happens, cuz urs doesnt even have a heatsink intsalled.

Hey, thanks for the reply I did not use a dedicated SSD for the OS, I will try that. I'm running Chia on a few other machines (intel NUC rigs), and for some of them, I installed the OS directly on a partition of the same SSD used for Chia plotting. Those machines are working fine without issue, so I doubt that's the issue in this case since it's the exact same setup with the SSD partition.

The second part of your advice I will also try today. I did not try a full cmos reset yet, hopefully that will fix the bios because before I could not even access the advanced OC settings without bios crashing/computer restarting.

Thanks for the advice about the SSD heatsink too, I didn't know it was necessary. I'll order one!