[Long] **UPDATE** Luminar keeps shutting down my PC. I finally seemed to have found the Culprit!! But now, I need a solution to this new problem. It’s CPU overclocking, confirmed.

[The original thread](https://old.reddit.com/r/overclocking/comments/170li4c/long_a_postprocessing_app_that_i_work_with_keeps/)

Here’s how I reached that conclusion all by myself:

So after setting up a tripod and opening up Luminar, I set up a tripod and put my phone up to record. The devs of Luminar told me to see which function inside the app causes the crash. I told them that it was not set in stone and it really wasn’t. Sometimes it crashed on the Denoise, sometimes it crashed on the sharpening, sometimes this, sometimes that. And so, out of curiousity, I set up the tripod to check. But my curiosity peaked even further. I opened OpenhardwareMonitor and opened only the metrics which I thought were important and still, it was like 20 metrics that were on screen.

For 20 minutes, I was testing various tools ranging from your run-of-the-mill editing tools, to extremes like the AI tools.

**Crash!**

I immediately took notice that it was the sharpening tool that I was just finishing up with. Then I quickly reviewed the footage. In a sea of 20 metrics, it was hard to pinpoint what was changing. Not a lot was changing so I had to spend 10 mins just scanning the last few moments.

One thing caught my eye – CPU clock on all cores. The maximum recorded was 4614 Mhz. My CPU is Ryzen 5700X which can overclock to 4.6 GHz according to the manufacturer. And I noticed that after I click on the sharpening tool to sharpen, the CPU cores (not all, but majority), drop to 3.6 which is idle speed, but then RIGHT BEFORE CRASHING JUMPING UP TO 4.61 GHZ THEN CRASHING. Ahhhhh. I think I know what this is!

Here’s how I tested my theory:

I searched how to overclock my CPU without manually doing it. People online were like you have to activate PBO first. Ok, then. I go to BIOS, find PBO, and it’s on advanced with some features on and others off. So I enable it which means turning everything to on. Boot up. The fans are already whirring louder. Good. Check Coretemp (CPU monitoring) and CPU is at 4.6 and temps around 65 – 70 deg C.

I go back to the search and people say that you have to stress test it. So I find Prime95. I put the test on High stress (not extreme – because I have a hunch that it’s going to do something spectacularly wrong in the first few seconds). Start test. 2 seconds later,

**CRASH!**

I FUCKING KNEW IT! But I need one more confirmation. I reckon it MUST BE stability. I open OCCT. Stability test. Large dataset (should be easier on the CPU). Run it.

**CRASH!**

Awesome! Thanks to me, I found the culprit!

***Here’s where I humbly ask for your help***:

I upgraded this computer approx 8 months ago and the parts that I upgraded are exactly the parts that directly involve this issue:
PSU, CPU, CPU cooler, and an extra M.2 SSD.

**Update:** TL;DR It’s the PBO and Curve Optimizer that is the problem.

Step 1: I turned off PBO (from BIOS) and Curve optimizer and ran OCCT Large, Medium and Small. All okay.

Then I am currently running Prime95 on small – high (then it is Small – extreme) and it’s going well so far. Though, on OCCT it did spike up to 4.5 GHz, but Prime95 is constantly at 3.6 GHz. Prime95 ran tests for a bit but did not crash and I stopped it. It was stuck at 3.6 Ghz.

Step 2: I went back to OCCT, reset the clock speeds, opened OHM as backup and reset speeds, set data set to Small, test duration 5 mins, and set it to clock cycling and for that 8 seconds cycling testing all 16 threads. All good.

Step 3: Ok. XMP enabled. I ran OCCT again and it works fine. That’s strange. I think it’s either the curve optimization or the PBO.

Running prime95 now and it didn’t crash 2 seconds in. Hmm…

Step 4: Someone suspected that it could be Curve Optimizer and PBO, so they suggested that I turn them on and put CO value at -5 on all cores. I just checked the advanced tab and the default value for all core is set to -14. When I set the CO manually to -5 cores, and ran OCCT’s hardest stability CPU test, it crashed 2 seconds in.

Step 5: I turned off CO and PBO and re-ran OCCT’s hardest test, and it passed again. So it’s PBO and CO’s fault.

Step 6 (ongoing): I am looking at some previous threads posted on reddit and other forums about PBO and CO and it turns out that it’s pretty common for PBO and CO on auto causing crashes, so I will have to set it manually using this video https://www.youtube.com/watch?v=dU5qLJqTSAc