CPU & RAM stress test that detects corrected ECC errors?

High-Performance Computing Workstations & Servers
1

I am using [Advanced Clustering Breakin][1] to stress test my Opteron 6276 & currently 2x8GB of PC3L-10600R on my SuperMicro H8SGL-F and I am getting the following errors followed by numerous corrected ECC Errors after only a few minutes…

ERR 00h 03m 37s Jul 15 07:59:57 (none) user.emerg kernel: [  299.816039] [Hardware Error]: CPU:0    MC4_STATUS[Over|CE|MiscV|-|AddrV|-|-|CECC]: 0xdc104000ab080a13
ERR 00h 03m 37s Jul 15 07:59:57 (none) user.emerg kernel: [  299.816329] [Hardware Error]:     MC4_ADDR: 0x000000014ec1e610
ERR 00h 03m 37s Jul 15 07:59:57 (none) user.emerg kernel: [  299.816406] [Hardware Error]: Northbridge Error (node 0): DRAM ECC error detected on the NB.
ERR 00h 03m 37s Jul 15 07:59:57 (none) user.err kernel: [  299.816554] EDAC amd64 MC0: CE ERROR_ADDRESS= 0x14ec1e610
ERR 00h 03m 37s Jul 15 07:59:57 (none) user.emerg kernel: [  299.816575] [Hardware Error]: cache level: L3/GEN, mem/io: MEM, mem-tx: RD, part-proc: RES (no timeout)
ERR 00h 03m 45s ecc - mc0:csrow2 error counter at 1
ERR 00h 03m 45s ecc - mc0:csrow3 error counter at 0
ERR 00h 03m 56s ecc - mc0:csrow2 error counter at 1
ERR 00h 03m 56s ecc - mc0:csrow3 error counter at 0
ERR 00h 03m 57s failid - Other tests have errors, turning on ID light and exiting.
MSG 00h 03m 57s Disabling burnin test 'failid'

Does anyone know of another program that can stress the CPU and Memory at the same time while being able to detect corrected ECC errors? I am currently not 100% sure if these errors are a defective CPU or the linux kernal having a seizure. Also I know this is not a RAM problem. I have already tested several sticks in various configurations. Always get the same error every time. CPU temps are not a problem and latest bios, etc. I am almost certain it is either the CPU or a stress test bug. It would clear things up if I could confirm or exclude false positives.

Lastly any advice on BIOS settings? I tried turning numerous things on and off and it does not seem to help but I may have missed something.

Edit: Updated log to have more detail.
[1]: Breakin Hardware Diagnostics - Advanced Clustering Technologies

2

These look like corrected ECC errors. Do you have any options to play with timings or dram refresh cycle? Sometimes you need to change the refresh cycle to be more often otherwise it manifests as

3

Hi,
I am getting same errors after 2 hours running breakin, i did memtest for 8 hours and didnt show any errors. Could you please tell me if you solved this issue ?