Some CPU temperature over threshold was considered fail but some wasn't

[Eric Fu]

Hi Sirs,

We have two test units with the same H/W and S/W configurations. One unit was reported fail in CPU temperature over threshold. The other wasn't but actually there were some CPU temperature over threshold. We need to figure out why some CPU temperature over threshold can be considered fail but some can be considered pass.

BurnInTest V9.2 1003 (64-bit)

System summary:
Windows 10 Professional Edition build 19042 (64-bit),
1 x 11th Gen Intel(R) Core(TM) i3-1115G4E @ 3.00GHz [2189.0 MHz],
7.7GB RAM,
Intel(R) UHD Graphics,
119GB SSD,

No.1 test unit
******************
DETAILED EVENT LOG
******************
CRITICAL: 2021-06-07 02:01:47, Temperature, Temperature threshold exceeded
LOG NOTE: 2021-06-07 02:01:47, Temperature CPU 0 core 0: Current 91.0C / Threshold 90.0C
CRITICAL: 2021-06-07 02:04:07, Temperature, Temperature threshold exceeded
LOG NOTE: 2021-06-07 02:04:07, Temperature CPU 0 core 0: Current 92.0C / Threshold 90.0C
CRITICAL: 2021-06-07 02:07:17, Temperature, Temperature threshold exceeded
LOG NOTE: 2021-06-07 02:07:17, Temperature CPU 0 average: Current 91.0C / Threshold 90.0C
CRITICAL: 2021-06-07 02:07:17, Temperature, Temperature threshold exceeded
LOG NOTE: 2021-06-07 02:07:17, Temperature CPU 0 core 0: Current 91.0C / Threshold 90.0C
CRITICAL: 2021-06-07 02:08:28, Temperature, Temperature threshold exceeded
LOG NOTE: 2021-06-07 02:08:28, Temperature CPU 0 core 0: Current 92.0C / Threshold 90.0C
LOG NOTE: 2021-06-07 02:11:40, Status, Summary results written to log file (2)

****************************************
RESULT SUMMARY - INTERIM PERIODIC REPORT
****************************************
Test Start time: Mon Jun 7 01:11:37 2021
Current time: Mon Jun 7 02:11:40 2021
Test Duration: 001h 00m 03s
Temperature CPU 0 average (Min/Current/Max): 40.5C / 81.5C / 92.5C
Temperature CPU 0 core 0 (Min/Current/Max): 40.0C / 83.0C / 92.0C
Temperature CPU 0 core 1 (Min/Current/Max): 41.0C / 80.0C / 93.0C

Test Name Cycles Operations Result Errors Last Error
CPU 229 5.658 Trillion PASS 0 No errors
Memory (RAM) 370 161 Billion PASS 0 No errors
2D Graphics 141 38311 PASS 0 No errors
3D Graphics 2 63773 PASS 0 No errors
Temperature - - FAIL 5 Temperature threshold exceeded
Sound 178 43.295 Million PASS 0 No errors
Video Playback 207 2651 PASS 0 No errors


No.2 test unit
****************************************
RESULT SUMMARY - INTERIM PERIODIC REPORT
****************************************
Test Start time: Tue Jun 8 01:02:05 2021
Current time: Tue Jun 8 12:32:07 2021
Test Duration: 011h 30m 02s
Temperature CPU 0 average (Min/Current/Max): 32.5C / 75.5C / 91.5C
Temperature CPU 0 core 0 (Min/Current/Max): 32.0C / 72.0C / 88.0C
Temperature CPU 0 core 1 (Min/Current/Max): 33.0C / 79.0C / 95.0C

Test Name Cycles Operations Result Errors Last Error
CPU 2639 66.541 Trillion PASS 0 No errors
Memory (RAM) 4429 1.865 Trillion PASS 0 No errors
2D Graphics 1630 440185 PASS 0 No errors
3D Graphics 23 694937 PASS 0 No errors
Temperature - - PASS 0 No errors
Sound 2053 498 Million PASS 0 No errors
Video Playback 2385 30529 PASS 0 No errors


Thanks,
Eric Fu

[David (PassMark)]

Unit 1 had both cores are nearly the same temperature. So the average would be very close to the individual core temp.
Unit 2 had a bigger difference. So maybe the average never got over the threshold, even if 1 core did for a period?

[Eric Fu]

Hi David,

I'm very confused about the over-temperature event mechanism.
We had another core i7 unit and its test results below show each core had ever encounter over 90C and even 100C in core 0, but no any temperature error was reported for the entire test.
Please give me more your comment or suggestion.

******************
DETAILED EVENT LOG
******************
LOG NOTE: 2021-06-18 05:03:27, Status, Test run stopped

**************
RESULT SUMMARY
**************
Test Start time: Thu Jun 17 17:03:14 2021
Test Stop time: Fri Jun 18 05:03:20 2021
Test Duration: 012h 00m 06s
Temperature CPU 0 average (Min/Current/Max): 72.5C / 77.0C / 95.3C
Temperature CPU 0 core 0 (Min/Current/Max): 73.0C / 77.0C / 100.0C
Temperature CPU 0 core 1 (Min/Current/Max): 72.0C / 76.0C / 93.0C
Temperature CPU 0 core 2 (Min/Current/Max): 71.0C / 79.0C / 93.0C
Temperature CPU 0 core 3 (Min/Current/Max): 74.0C / 76.0C / 95.0C

Test Name Cycles Operations Result Errors Last Error
CPU 2760 145 Trillion PASS 0 No errors
Memory (RAM) 4538 1.909 Trillion PASS 0 No errors
2D Graphics 1695 457661 PASS 0 No errors
3D Graphics 24 731015 PASS 0 No errors
Temperature - - PASS 0 No errors
Sound 2141 519 Million PASS 0 No errors
Video Playback 2488 31856 PASS 0 No errors
TEST RUN PASSED

Thanks,
Eric Fu

[Tim (PassMark)]

What sources did you have checked for the temperature monitoring or are you using the "Automatically select" option?

What threshold did you set (I assume 90C)?

Do you have a trace log from one of these test runs or if not could you please enable the trace logging level 2 in the logging options and send us the trace log.

[Eric Fu]

Sorry, I found the reason of the discrepancy. Our tester set the threshold as 110C for i5 and i7 CPUs, while 90C for i3 CPUs.

[David (PassMark)]

That would explain it.