OCCT's "Power" test CPU errors found

[newconroer]

A quickie hopefully
I have struggled to find any posts that are centered around the Power test or the above errors
---

Was working through new build
Did power test with OCCT with everything stock

Am I to understand that the Power test on OCCT incorporates CPU and GPU
As a result then should the issue be CPU (voltage for example) then I'd find core errors simultaneously under the CPU test?

Currently (at least when stock) I do not get any core errors in the CPU test

This might indicate the issue is related to the GPU
With that in mind I went to testing Blender, Valley/Heaven and some real games for a while - no errors crashes etc.

 

[3x0]

Currently (at least when stock) I do not get any core errors in the CPU test

CPU test with SSE enabled?

 

[MrDweezil]

Maybe running up against the limits of your power supply when maxing both your cpu and gpu? Didn't see the full parts list anywhere when skimming your linked thread.

 

[newconroer]

Maybe running up against the limits of your power supply when maxing both your cpu and gpu? Didn't see the full parts list anywhere when skimming your linked thread.

System spec if that helps (think I updated it before posting this thread)
I suppose that's part of the question. I can see clearly that it is raising CPU and GPU wattage during the test, though is that specifically what the power is aiming for, or is there more to it.

Last attempt CPU was 112w and GPU 240w
Not exactly burning the house down

CPU test with SSE enabled?

Ya had some runs with SSE and AVX2

 

[MrDweezil]

System spec if that helps (think I updated it before posting this thread)
I suppose that's part of the question. I can see clearly that it is raising CPU and GPU wattage during the test, though is that specifically what the power is aiming for, or is there more to it.

Last attempt CPU was 112w and GPU 240w
Not exactly burning the house down






Ya had some runs with SSE and AVX2

I see now. Yeah your PSU is fine.

 

[hat]

Maybe the power test uses a different CPU test than you do when you test the CPU alone? The power test probably uses Linpack, which would be more intensive than a standard CPU test. Could be a case of AVX options being different across the two as well.

 

[newconroer]

Maybe the power test uses a different CPU test than you do when you test the CPU alone? The power test probably uses Linpack, which would be more intensive than a standard CPU test. Could be a case of AVX options being different across the two as well.

Maybe, though I have done Linpack testing without errors as well
Also the main concern is why it's showing the error on stock settings

 

[tabascosauz]

Core "stability" is hard to nail down on Ryzen because of how it works. Bottom line, cores should never be erroring out at stock.

You can use some of the common curve optimizer testing tools to double check. The corecycler script, Y-cruncher stress test function, etc. I've never really been a fan of OCCT because it's not particularly transparent as to what it actually tests.

y-cruncher - A Multi-Threaded Pi Program

GitHub - sp00n/corecycler: Script to test single core stability, e.g. for PBO & Curve Optimizer on AMD Ryzen or overclocking/undervolting on Intel processors

Script to test single core stability, e.g. for PBO & Curve Optimizer on AMD Ryzen or overclocking/undervolting on Intel processors - sp00n/corecycler

github.com

 

[newconroer]

Core "stability" is hard to nail down on Ryzen because of how it works. Bottom line, cores should never be erroring out at stock.

You can use some of the common curve optimizer testing tools to double check. The corecycler script, Y-cruncher stress test function, etc. I've never really been a fan of OCCT because it's not particularly transparent as to what it actually tests.

y-cruncher - A Multi-Threaded Pi Program

GitHub - sp00n/corecycler: Script to test single core stability, e.g. for PBO & Curve Optimizer on AMD Ryzen or overclocking/undervolting on Intel processors

Script to test single core stability, e.g. for PBO & Curve Optimizer on AMD Ryzen or overclocking/undervolting on Intel processors - sp00n/corecycler

github.com

Thanks
Reading through the CoreCycler instructions, trying to find the actual settings/section to modify in the config.default, would like to change the time per core and some other things

 

[tabascosauz]

Thanks
Reading through the CoreCycler instructions, trying to find the actual settings/section to modify in the config.default, would like to change the time per core and some other things

You don't have to change the config. The default config is good enough for a quick test, just to make sure only OCCT has lost its mind. Changing the config is more for long-term curve optimizer testing where being on the edge of instability can be very unpredictable and elusive. At stock you should be well within stability.

Ycruncher is quite nice because the stress testing suite by default gives you a variety of tests that cover a wide range of loads ranging from core-heavy to uncore-heavy (Fabric and UMC).

 

[newconroer]

I do mean to have a curve optimizer arrangement, though hopefully the default for CoreCycler is sufficient
Will do the Ycruncher all component test as well

Pushed my curve settings pretty low - more so than OCCT will dare tolerate, and Corecycler has been pretty tame though I did have one system crash and reset(meaning it reset on it's own)
Looking at the log for CoreCycler itself, there's no listed errors - rather the logging seems to have stopped right as it was moving onto a new core
Same holds true for the PRime95 log - last thing listed was the previously passed test


This coming after about an hour of testing without issue

In a real world day to day way only 'crashing' I am getting is when in Warcraft where the condition is, the game has been idling some time and I have Edge browser open. The system will freeze up
Event viewer has no entries of note, sans this one

Autopilot.dll WIL error was reported.
HRESULT: 0x80070491
File: onecoreuap\admin\moderndeployment\autopilot\dll\dllmain.cpp, line 128
Message: NULL

When you look up what that .dll is, doesn't seem clearly related though there are reports that failed overclocking (e.g. bad configuration, voltage shortage) can cause it

Of course that could be wrong in my case but I don't know how to fix it (or what the intended fix is in normal circumstances) to be able to 'fix' it then see whether it keeps happening
normal circumstances) to be able to 'fix' it then see whether it keeps happening


So with all the above said, makes me wonder if it is purely down to too much of a negative Curve causing low CPU voltage specifically and core crashing, or the system overall is affected by it.