Windows 10: Exporting in Adobe Prem 11 Elements causes WHEA_UNCORRECTABLE_ERROR

After checking the connections to the CPU fan on the motherboard, it appears as though some connections were reversed.
The CPU fan was connected to a pinout called "CPU OPT", while the "radiator" fan (the part of the water cooler attached to the side of the box) was plugged into the "CPU FAN" pinout next to it. Reversing these only slightly changed the temperature properties. I am going to reapply the paste, since I believe it wasn't applied properly last time.

Paul

 

To answer your question, I have not yet had another BSOD, and the 100C issue seems to arise most when I use the "Small FFT" setting, which generates the most heat, reaching 100C in under 5 minutes. Under the default test ("Blend"), I never get to 100C, although I do get as high as 96C, but stay at about 75C if the load is said to be 99.9%. Right now, the load is 99.5% and the temperature is below 72C most of the time, which agrees with the CPU specs. This test has been running for 15 minutes.

After 17 minutes, 2 errors were observed in two different virtual cores and I stopped testing. The CPU was overheating at any rate at the time.

 

Meaning the CPU is likely damaged.

 

I hope a damaged cpu is not the case, but something isn't right.

I ran memtest86, which reported 19 errors in specific areas located on one of the DIMMs after 4h (CPUs running in parallel) (tests 6-9), but this appears to be not the case (0 errors after 2 passes)after running the same test on the same memory locations the errors were reported in (0x2039f007c to 0x3e8f0fe18) when the CPUs are running in sequence.

The docs to Prime95 tell me that another possibility is a power supply that is not up to the level of the other hardware. But if it is the CPU, then I don't want to just replace it thinking the problem is solved. I would still like to know what is causing the flaky CPU. The pinouts for the fan, even though they were reversed, were the same kind of connections. CPU FAN and CPU OPT both operate fans and have the same pin connections and voltages.

 

Please run MemTest86+ again for at least 8 complete passes, make a photo of the result and post the photo. This to make sure it is properly done.

Check the voltages of the componenten and the PSU.

 

Memtest won't allow me to do more than 4 passes at a time. I have version 6.1.0 (Free) from their website, which was burned onto a CD. I can post 2 photos of 4 passes (4h long with the limited amount of memory I am testing now). After 3:10h, no errors.

When I do the second run, I will test all of the RAM with the sequential setting, and I will remember to look at the voltages when I reboot.

 

Please use version 5.01 or 4.20

 

The UEFI doesn't tell me the voltage being passed to the PSU. (power supply)
Other voltages which it monitors:
All 4 cores: +0.992V
The rest seem to vary a tiny bit, but here are some typical values:
CPU Cache: +1.094V
CPU System Agent: +0.824V
CPU Analog I/O Voltage: +1.012V
CPU Digital I/O Voltage: +1.021V
CPU Input Voltage: +1.696V
DRAM Voltage: +1.535V
PCH Voltage: +1.059V
3V pinout: ~3V (varies in small amounts)
5V pinout: ~5V (varies in small amounts)
12V pinout: 12V (exact, no change)

 

I am currently running Memtest86+ 5.01, as suggested, and I am a bit concerned that there is a line that says:
Cores: 1 Active: 1 / 1 total (Run: All)

I in fact have 4 cores, and the top line on the screen identifies it correctly as an Intel Core i7-4770k

It is still running as it has completed only 1 pass so far.

Screenshot:
Attachment 36290

 

0 errors after 8 passes, and just under 35 hours of running time

screenshot:
Attachment 36519

 

Did another video rendering using Adobe Premiere Elements, and this time, no BSOD. Also, your next test for testing my C: drive (an SSD) turned up nothing under CHKDSK.

 

I also made improvements with fan placement and placement of the CPU fan/radiator assembly in the box, and moved some cabling out of the way to improve airflow.

As I was cleaning the fins on the radiator, I noticed that hardly any dust had accumulated on it or on the screen that covered the fan. When plugged in, however, the fan was spinning.

Prime95 now runs at 91C or less under 100% CPU load. An incremental improvement, but still not out of the woods. My idling temperature (as in right now as I am writing this) is 25-28C, a noticeable improvement.

I also read/saw a thread and a video posted to overclock.net regarding a known problem with the Intel Core i7 4770K and Haswell processors generally. Apparently, the heat transfer between the CPU and the heat spreader is not that great. There are people, such as the person in the video going so far as to adding thermal paste underneath the heat spreader (beteen the HS and the CPU itself). The fellow in that video in the "before" scenario has exactly the same heat problems as me. The "After" video showed his processor running Prime95 at under 70C.

I am not sure if you actually meant to add thermal paste under the heat spreader ... ? *Rolleyes