PC Processors

[agatek]

Hi,

I am facing some restarts issues with the above said cpu on Asus PRIME X370-PRO, most recent bios (6042). No overclocking etc.

Exactly the same hardware had no such issues running 5-3600X i 7-1700PRO.

The restarts occur when executing any more complex software both under W10 and Linux (Mint). I don't know too much about Windows but for every restart under Linux this errors appears in the kernel log:

[ 316.856151] [Hardware Error]: Corrected error, no action required.
[ 316.856156] [Hardware Error]: CPU:0 (19:21:2) MC12_STATUS[Over|CE|MiscV|AddrV|-|-|SyndV|CECC|-|-|-]: 0xdc2040000602010b
[ 316.856160] [Hardware Error]: Error Addr: 0x00000000000a9f40
[ 316.856161] [Hardware Error]: IPID: 0x000700b020350500, Syndrome: 0x000000232a1f0f0e
[ 316.856164] [Hardware Error]: L3 Cache Ext. Error Code: 2, L3M Tag ECC Error.
[ 316.856166] [Hardware Error]: cache level: L3/GEN, tx: GEN, mem-tx: GEN

There is not much info googles throw out when looking up "L3 Cache Ext. Error Code: 2, L3M Tag ECC Error" but one of the links describes pretty identical behaviour with the final outcome: cpu was defective and the problems stopped after changing it.

What may I still want to check before sending the cpu to AMD?

agatek

[heat]

Did you ever sort this out? Not a lot of info on Google about this..

I got something similar on a 5600X on a B550 MSI mobo

[Hardware Error]: Corrected error, no action required.
[Hardware Error]: CPU:1 (19:21:2) MC11_STATUS[-|CE|-|-|-|-|-|-|-]: 0x8000000577f15163
[Hardware Error]: IPID: 0x0000000000000000
[Hardware Error]: L3 Cache Ext. Error Code: 49
[Hardware Error]: cache level: L3/GEN, tx: INSN

[misterj]

agatek, I know almost nothing about Linux. What is the effect of the error - crash? Could you please run Windows and post a screenshot of the error in the Event viewer, basic view? I suggest you open an AMD Support request. Please post all your system components. Thanks and enjoy, John.

[agatek]

The mode of failure also under Windows was just random-like reboot with no error messages present on the screen.

[elstaci]

Open a AMD SUPPORT - Warranty and see if they believe you need to RMA your processor or not from here: https://www.amd.com/en/support/contact-email-form

In Windows you can download freeware OCCT and stress test your CPU, GPU, & PSU to see if it passes all the tests without stopping or shutting down or BSODing.

Keep an eye on Temperatures, Fan Speeds, & PSU Outputs for any abnormal readings.

NOTE: ECC Error seems to indicate your RAM. Do you have ECC RAM Memory modules installed  by any chance?

You might also want to physically check your Ram for being defective by either running MEMTEST64 using a Flash Drive or Windows Memory Diagnostic app.

[agatek]

The RAM was non-ECC, but the error in the logs suggests problems within the L3 cache. L3 cache is ECC and a part of CPU.

Regardless, I actually tested RAM with the memtest. Stress tests were also performed with dedicated software under Linux (stress, stressng). Interestingly, what rather quickly led to reboots under Linux were not the stress tests but  ffmpeg where cpu saturation was lower than in the stress tests.

[elstaci]

Thanks for the explanation concerning ECC which I wasn't aware of. Good to know for the future.

[agatek]

Shortly after posting the initial message I opened the process for RMA and all seemed to be heading towards a successful return but at one point AMD support mentioned it could be more straight forward and less expensive to return it via the supplying retailer. I bought it from Amazon but it was already a few month after free return window, still I contacted Amazon CS and they agreed to accept the return (as it concerned a defective item) warning me at the same time there would be some deduction (for being late). As prices for the cpu between time I bought it and returned dropped rather significantly I actually replaced the cpu with 7-5700X, inserted to the same hardware and all problems disappeared.

So the story with happy ending and the technical conclusion pointing to a faulty cpu.

[elstaci]

Hey Thanks for the update.

Generally most processor are sturdy and can take a beating and work for years without failure unless it was defective from the factory.

AMD Warranty should have been in effect since Amazon Return date or Warranty had expired already. But if you were charged a small late fee for returning a DOA processor then Kudos to Amazon for their great return policies and Customer Service.

How long that will last, concerning Amazon, if they start losing a lot money or business in the future that will change.

Glad you got your PC  working normally again.