New build pc build, lots going wrong

[MrAlfredBuddyGuy]

I’m trying to build a new pc for my brother who’s been using a 10 year old hp for the longest time.
OLD PC was an office handme down, maybe 10 years old . I7 2000 series I think. He had an RX 580 and all he did was play pubg on 720p because he uses a really crappy tv and refuses to buy a monitor.

PC started crashing non stop, and frankly being tipsy often led to smashed USB ports and even him killing one of his SSD’s.

I’ve been telling him he should upgrade for a while now and so he finally hit me up to start a new build.

We have a Ryzen 5 5600
16GB Ram
500GB SSD
RX 6600XT

I’ve been having the most difficult time though. Not my first rodeo, built several pcs for myself and work.

I created a windows USB boot drive with rufus tool, often it would crash during installation, with generic windows blue screens. Even during install from the USB drive it would say someone about needing drivers and it wouldn’t actually let me install windows.

Decided to download the windows iso again and use Rufus to make another bootable image and it finally installed.

Things looked good but after I left her said he would crash and get blue screens.

The most notable is something like IRQ and if I remember correctly that is usually a driver issue.

So I told him I’d take it home and run some tests.
Prime95 wont run “hardware failure detected running 960k FFT size, consult stress.txt file.”

Super position runs but after the unigen splash screen and the chalk board looking screen the screen is black, days benchmarking at the top asking with performance metrics on the right.

I don’t know where to go from here. The only thing that isn’t new is the power supply but it’s only a few years old, it’s an evga 600watt.

screenshot he took

Prime 95 results
[Fri Mar 17 21:16:05 2023]
FATAL ERROR: Rounding was 0.4983563445, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.
FATAL ERROR: Rounding was 0.4995133954, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.
FATAL ERROR: Rounding was 0.499821732, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.
FATAL ERROR: Rounding was 0.4998055169, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.
FATAL ERROR: Rounding was 0.4973511869, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.
[Fri Mar 17 21:18:52 2023]
FATAL ERROR: Rounding was 0.4999710343, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.
FATAL ERROR: Rounding was 0.4994222756, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.
FATAL ERROR: Rounding was 0.4991478204, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.
FATAL ERROR: Rounding was 0.4999706614, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.
FATAL ERROR: Rounding was 0.499584915, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.
FATAL ERROR: Rounding was 0.4989862138, expected less than 0.4
Hardware failure detected running 960K FFT size, consult stress.txt file.

[regulareel]

Have you tried reseating the CPU and RAM?

Also can you copy and paste the actual contents of stress.txt?

[MrAlfredBuddyGuy]

post edited with prime test. Haven’t done any re seating yet. Was planning to get memtest on a flash drive though soon too potentially.

But I can do a re seat in the morning. Winding down for the night.

[jode]

Temps?

[MrAlfredBuddyGuy]

prime 95 this morning, seems like 1 core is working without failing… haven’t reseatted cpu yet, just memory.

super position still “runs” but the screen is blank and I know it shouldn’t be obviously. but if I minimize I can see that gpu load was max in task manager, but as soon as I try to click on the icon in task manager to resume, it crashes.

[MrAlfredBuddyGuy]

run super while I was in the shower for about 10 minutes ( and again just a black screen and it would never change from scene 1) but here are the temps from that. But they seem normal to me.

[wUFr]

idk how picky are 5th gen. ryzens on RAM, but maybe try lowering its speed? Running XMP profile?

Temperatures on some of those PINs on board are “a bit” toasty, but i assume its just additional connector for optinal sensor? if not, something is definetelly cooking over there, maybe open it and actually check if there isnt something toasty

[MrAlfredBuddyGuy]

@wUFr

lly check if there isnt something t

On second glance, the only thing that looks hot is TMPIN6 and 8 on the mobo. but I don’t know what those are. Is that what you were talking about?

I’ll look into memory profiles, but at the moment everything is stock.

Also I experienced that AMD Bug Report window myself just now loading up youtube on microsoft edge. screen went totally black for a few seconds before it popped up.

[wUFr]

yes, id really open the case and put hand near VRM heatsink to check if thats not cooking itself to instability.

…tried to check mobo manual, but couldnt find any specific info about temperature sensors tho (which is which)

another way would be checking temps in bios, if they are under 40°c on idle

---

Also i think there is/was some serious issue with windows/amd driver which would screw itself and system as well in the update process, may wanna try full driver wipe and windows repair command.

[MrAlfredBuddyGuy]

I mean, this is literally a clean install of windows from the get go (only a week old at this point), plus ninite for some basics, firefox, steam vlc. I downloaded amd driver directly and didn’t rely on windows update.

I am going to swap his gpu into my pc and see if I have any issues there, I guess that’s also one way to rule out a potential hardware problem since potentially this system may have multiple issues.

[wUFr]

the issue appeared in manual update if you checked “factory reset” checkbox, not windows update, but ye, trying GPU in other rig is a good start

timestamped: HW News - AMD RX 7800 Leak, GN Imposter, Intel 1.8nm CPUs, Arc Even Better - YouTube

probably not the issue, but may give a clean driver install and windows verification a try

[MrAlfredBuddyGuy]

Alright so, I moved the gpu into my PC. Ran ddu and installed clean drivers. Tried to run super position and I get the exact same problems.

The benchmark pretends to run, but it’s just a black screen. Then if you minimize and come back, the closes the benchmark window.

Any chance the PSU could have damaged components, since it’s the only thing that isn’t new? I have a hard time believing that would be the case but I wouldn’t even know how to check.

If I remember correctly Ryzen. 5000 series doesn’t have on board graphics. And at this point I would be scared to try my gpu in his system… Cause actually this gpu in his previous system died and I’m paranoid.

Originally the plan was to reuse his old gpu, but it was a no go from the start, simply no picture. I installed it in my home office PC and no picture either. Installed it on my home PC and it was no bueno.

[wUFr]

if it uses same PSU as the rig where his old GPU died, thats a bit sus… guess RMA the new graphics card and buy new PSU to be sure, espetially if the old PSU is something lower tier

[MrAlfredBuddyGuy]

I mean it is kinda low tier, but 600 watts should be plenty for his system.

Also, going back to the TMPIN things in HWMonitor, I was reading similar complaints on reddit but some people were saying that those may just be sensors that don’t exist. I put my finger on the vrm heatsink and it wasn’t too hot to touch. and if it really was 100c there’s no way I could touch it for more than a second without wanting to pull away.

Also, even though super position was not giving me good results. I tried playing diablo 4 as well on my pc, and it stutters. So just more confirmation something is not right with the GPU.

[wUFr]

ye it could be unused sensor, some boards have additional headers to plug sensors in too.

“600W” doesnt really mean anything these days tho, id just buy new one at this point, it could just have some weird power spikes that killed something on the gpu (one YT video mentioned these are usually related to bad memory, doesnt really matter at this point i guess)

maaaaaybe you could try downclock the gpu/vram

[MrAlfredBuddyGuy]

I connected my power supply to his components. Still can’t run prime95 for very long before getting the same errors I reported earlier. at first it looked like it only errored immediately on 1 core and the other 5 kept going but seconds later 4 more cores had errors also.

[anon7678104]

looking at the prime95 error its the ram.
its either bad or running bad timings.

you could try updating the bios to the latest memory compatibility upgrade.
may fix it.
or adjust the timings.

[wUFr]

he has GPU giving him artefacts in a different PC. Im wondering if both ram and vram could get damaged by a bad PSU, cuz that would suck.

[GigaBusterEXE]

Run the ram in another system and run a memtest on it, windows has one built in just type memtest in the start menu and it’ll reboot into it, takes a while though

If it passes ram itself is good, your GPU is probably bad, but that doesn’t mean everything else is out of the woods, bad mobo could give memory errors
But so can a CPU so better work it out quick while you still have a warranty

[MrAlfredBuddyGuy]

I’ve put his memory in 2 other sytems.
A Ryzen 3600x and a 3700x.

On the 3600x, prime95 failed with those same kind of errors on most of the workers (I didn’t give it more than about a few minutes)

On the 3700x, it’s been about 10 minutes and half of the workers have failed.

Ran mem test on the 3700x and it failed. So my last test will be putting good memory in the new system and trying prime95 again.