Question Constant BSOD - multiple errors

[splendy]

Hey guys! Hoping one of you can help me out.
Last 3-4 weeks my computer has had an increasing amount of crashes in all games, especially unityengine games, and now i'm starting to get BSOD if i'm just browsing the net too.
I've tried multiple windows reinstalls, including a complete fresh one yesterday.
I've tried updating every driver/reflashed my bios/using only minimal drivers
I've memtested my ram using memtest86 for 3 passes (both sticks plugged in no fails so didnt bother to check them individually or change their ports)
I've bought a new PSU as I thought it was power issues and then after that bought myself a new gfx card yesteday, in which the problems STILL occured and so I reinstalled windows completely (fresh wiped both drives), crashing still occuring.
Running out of options
posting now and edit it later incase I BSOD while posting it
gonna post my crash logs here:

https://drive.google.com/file/d/1kuRZzpI9o_yKUOHJOi7AbJwMYmcS7vSG/view?usp=sharing

https://drive.google.com/file/d/1rGiwy3931ReV0kgjx34L78C00yt6glbo/view?usp=sharing

https://drive.google.com/file/d/1OyJZ1XA5nz2R15EgkxYziokaOAw5EHjm/view?usp=sharing


SPECS:
Processor: AMD Ryzen 5 3600 6-Core Processor (12 CPUs), ~3.6GHz
Card name: NVIDIA GeForce RTX 2060
MOBO: asrock x570 gaming phantom 4
MEMORY: 16gb (8x2) 3200 mhz ddr4
PSU:Coolermaster 600W Gold standard

 

[splendy]

https://drive.google.com/file/d/1NR0PWsBImeUvtXw4FwVjLOqV1GP4CXHr/view?usp=sharing

https://drive.google.com/file/d/1a6HcHnCUx3DQW3x-MIvp4rxslaVlM-ZC/view?usp=sharing

Here are some other bluescreen logs. I had unplugged my HDD and was just using my SSD as I was testing whether my SATA cables were broken, and this occured, making me think that my SSD (where my paging file is) might have become corrupted/broken.

 

[Colif]

i will ask a friend to read the dumps for us. He will reply here with a file I can use to read them.

what brand ssd is it?

could try running Prime 95, it will test CPU & RAM - https://www.mersenne.org/download/
Prime 95 how to Guide: http://www.playtool.com/pages/prime95/prime95.html

Since it survives reinstalls it is likely hardware
if you are getting dump files, the page file is working in some way as it is used to create the dump files.

 

[splendy]

Hi,
it is samsung evo 360. I've tried using SPtoolbox and another SSD type tool to actually test the SSD itself but both of them kept saying that the SSD was not able to be found. Samsung magician locates it fine though but only lets me benchmark it.
I'll try prime95 now, thank you!
EDIT:
Prime95 is giving me hardware failures on some of the workers, particularly 1 and 2 before the other 10 workers were able to actually crunch numbers.

 

[Colif]

360? typo? they don't have a 360... maybe an 860?

Prime95 is giving me hardware failures on some of the workers, particularly 1 and 2 before the other 10 workers were able to actually crunch numbers.

make sure its set up right, sometimes it fails if its not set up right

 

[splendy]

yes correct, type, 860 whoops! Another worker has failed, 3 workers failed out of 12, with my CPU running at full blast.
I will reconfigure and try again. I had used a blend test which kept my CPU and RAM both at 100% with 3 workers failing after 17 minutes, the first two workers instantly and one later.
I'm now using a Small FFTs test to test my CPU specifically, two workers again failed within 5 seconds, the other 10 grinding hard at the numbers.
A third failed at 2 minutes in.
Using HWinfo while i'm running this i'm noticing that my effective clock in core 0 is suspiciously low, which could be CPU's associated failing workers, 1 and 2. Could a core in my CPU be faulty? (due to ryzen 5 3600 threading, technically 2 cores)

3 workers both times dying almost immediately, while the other 9 are working fine. This is only over 15 minutes too. (workers 1,2 and 12 dying)

 

[Colif]

could be its the CPU and not the ssd.

Shame the only real test for Ryzen cpu is Prime 95

Prime 95 creates a report called stress.txt in the folder its installed in, might ask you to upload it to a file sharing website once tests all failed so I can see what failed.

No overclock.. base is 3.6 so I assume not.
is ram on xmp? try running it at stock and see if results change in Prime.

 

[splendy]

RAM is currently not using an XMP profile as I thought it was the stem of my unityplayer crashing issues in games. It is currently at boosted clock rate to 3200 (which is its intended mhz) but i'll put the bios back to the default 2400 (i think).

Will upload the fail log here:

 

[Colif]

3200 is its xmp speed, its counted as an overclock. 2400 is its base speed.

 

[splendy]

Well, when I switched the bios back to auto, which is by default 2400mhz the computer started fine. I have now crashed twice while running blend test with prime95

This is the error.txt from prime95 (the stress.txt never actually updated it just says default stuff but ill upload it anyways just incase)
https://drive.google.com/file/d/1DcNiuBYpt9zbbaGPAlzjtWvkfRlOFjNV/view?usp=sharing
(this is stress.txt)
https://drive.google.com/file/d/1rc81ow59jhW5K_gf_89AGDo9DjN5N-ME/view?usp=sharing

These are the crash logs minidumps from BSOD when I tried prime95 with 2400mhz ram.
first crash
https://drive.google.com/file/d/1w0ZLmbh0wW1KXiGCpGKEaDjFH9uoqrF2/view?usp=sharing
second crash
https://drive.google.com/file/d/1kc-j9dXcLHsoCKqEWPUHwiFyfUABhqr8/view?usp=sharing

 

[splendy]

Trying another small FFTs test instead of a blend (which doesnt seem to blue screen me)
Workers 1 and 2 both stopping within the first 60 seconds, causing core 0 to have very low clock ratios.
15 minutes in, only worker 1 and 2 both dead, rest is fine.
16 minutes in and worker 12 died. The same 3 workers keep dying every time I run it.

 

[Colif]

rounding errors can be fixed by giving ram more voltage

no point looking at dumps if ram is the problem, it effects everything run on pc.

 

[splendy]

do I give my ram more voltage in bios? or is this possibly a ram fault and I need new sticks
I gave my ram 1.350 volts compared to the usual 1.2 volts and the first two workers still died both times I ran it due to rounding errors.

 

[Colif]

faulty ram more likely than 3 cores on CPU not working.

You probably notice if it was the cpu. If your CPU anything like my 3600xt, core 0 is the most active so it would be obvious.

 

[splendy]

So even if i'm running prime95, and my core 0 has its effective clock VERY low (500-1500) while the other cores are at a stable 3900, you still think it's the RAM?
(bare in mind this is just effective clock, active clock still has the core 0 floating between 3000 - 3900 mhz while the other cores stay at a stable 3900~)
i'll run another memtest86 over night on my ram to double triple check it.
I think you might be right, I will buy another set of ram sticks and test them out comparatively.

 

[Colif]

Effective is more accurate at showing real speed. active just shows the base clock times clock speed, it doesn't account for sleeping cores.
Active is closer to what task manager thinks my CPU is always running at, effective is closer to what Ryzen Master shows it runs at. I use HWINFO to track effective.

right this moment, i don't know which to blame. memtest just tests for bad blocks but may not test if its running right. I haven't had someone fail core tests so fast before so I am hesitent to say which it is.

I would have gotten a shop to test parts way before I started replacing them. I would still do that at this point, just to save any more wrong guesses.

 

[splendy]

time to take it into the shop then