Question Help with diagnosing freezing/crashing when rendering certain graphical effects.

cseck28

Distinguished
Jul 21, 2015
6
0
18,510
0
Specs:
Operating System
Windows 10 64-bit
CPU
Intel Core i5 6600K @ 3.50GHz
RAM
16.0GB Dual-Channel DDR4 @ 1066MHz (15-15-15-36)
Motherboard
MSI Z170A SLI PLUS (MS-7998) (U3E1)
Graphics
8192MB ATI Radeon RX 480 Graphics (MSI)
Power Supply
EVGA 650GQ 80+ Gold

I just recently updated to Win 10 after holding out for years and ever since I've been having a problem when gaming where I will suddenly lose signal to the monitor and it won't recover until I manually power down the system and turn it back on. It doesn't appear to hard-freeze, at least not initially, because the lights on the keyboard for caps/num lock still turn on and off and I can still hear audio from the game and anything else I am listening to. After 30 seconds to a minute or so the computer does seem to freeze or crash completely. I figure this is either caused by my PSU or GPU, but not sure which.

My temps seem ok. My CPU is only around 30-40c and my GPU idles at around 50 and climbs up to around 75 under load so I don't think it's overheating. I never get any BSOD crash dumps, and event viewer only lists the unexpected manual shutdown as critical. I notice a pattern that rendering heavy dust or smoke effects in particular in these games seems to trigger the problem I'll list some examples of video (not mine) of things in-game that trigger the issue 100% of the time with timestamps. These are M rated games, so be warned of violence/language:

View: https://www.youtube.com/watch?v=C2kN94aFMcA&t=3m43s
3:43
28:28
59:05

View: https://www.youtube.com/watch?v=vRI4eKg9O8k&t=25m13s
25:13

Terminator Resistance is a relatively new an unoptimized game that doesn't run exceedingly well for me (can only get around a stable 30fps) so if it were the only game causing problems for me I'd maybe just chalk it up to some quirk between my hardware and that game, but I also get this issue during this particular cutscene of collapsing rocks plays in Call of Juarez: Bond in Blood, which is a 13 year old game and two (gaming) generations behind Terminator, and my PC should (and does except for this one instance) run it lightning fast

View: https://youtu.be/4c2zeZRvE7k?t=7400
2:03:22

I've also noticed the freeze/crash occasionally happens when there is a large on-screen explosion in various games like Sniper Elite 3 and Mad Max, which are also games I can otherwise run smoothly at 60fps with no problems. I can run the AMD radeon software stress test with no problems, but I also tried running furmark as a stress test and it IMMEDIATELY causes the problem as soon as I start the test. I also tried undervolting the GPU a little, didn't seem to help.

Possibly related: Ever since I build this PC around 2016 or so, I've also had an intermittent problem where the PC will hard freeze for seemingly no reason when idling or under low workload and never when playing a taxing game or anything like that. When this happens, I don't lose the video signal unless I turn the monitor off and on again, and once again no crash dump/relevant event viewer info. Switching OS from 7 to a fresh install of 10 didn't fix this problem so it must be hardware related. It only happens around once every few weeks or so, so it's something I've just tolerated all these years but I figure it could be related to my new problem.

I don't think the switch to Win 10 necessarily caused the problem even though that's when the issue started. (I also tried using older AMD drivers before they dropped support for 7 to see if that would fix the problem just because those were the ones I had been using before, it didn't.) What I'm wondering is if my GPU is just starting to die or my PSU isn't working correctly and causing things to crash/freeze. Unfortunately I don't have access to a spare PSU or GPU to test without paying for one so I'd just like some extra opinions before I start spending hundreds of dollars on replacement parts. Thank you for any help and advice.
 

A_Goat

Distinguished
Do you have XMP enabled in BIOS?
Are you able to reproduce the problem after resetting BIOS to default settings, with XMP off?

When you went to windows 10, how did you upgrade? Fresh install of windows and losing all your data, or upgrading 7 to 10?
I'd focus on some more troubleshooting before assuming it's hardware related, but right now I'm leaning more towards GPU especially since you can reproduce the problem easily with furmark.
 

A_Goat

Distinguished
XMP is off and it was a fresh install of 10, not an upgrade. The BIOS is already set to default settings.
Have you tried enabling XMP?

Sounds like GPU stability could be the issue.
Does your GPU sag at all in the PCI-E slot?
Tried blowing out the PCI-E slot to remove dust, or trying your other PCI-E slot?
Does your GPU have a VBIOS switch? I found mixed answers online for your specific GPU. If it has one, it will be a tiny little switch around the PCB to switch to a different VBIOS, if so, try flipping that switch. If not, I'd try updating the VBIOS manually.
 

cseck28

Distinguished
Jul 21, 2015
6
0
18,510
0
Enabling XMP doesn't seem to make a difference. I actually did try re-seating the GPU and blowing out dust with some air duster. I can try putting it in a different PCI-E slot a little later when I have some more time.

I don't think my GPU has a VBIOS switch. This is my exact model
https://www.amazon.com/MSI-GAMING-CrossFire-DirectX-Graphics/dp/B01K1JTT8S?th=1
I'm not sure how I'd go about updating the VBIOS. IS that something I would get from AMD or MSI?
 

A_Goat

Distinguished
Enabling XMP doesn't seem to make a difference. I actually did try re-seating the GPU and blowing out dust with some air duster. I can try putting it in a different PCI-E slot a little later when I have some more time.

I don't think my GPU has a VBIOS switch. This is my exact model
https://www.amazon.com/MSI-GAMING-CrossFire-DirectX-Graphics/dp/B01K1JTT8S?th=1
I'm not sure how I'd go about updating the VBIOS. IS that something I would get from AMD or MSI?
What about GPU sag? See picture. I ask because if it sags, as is common with old GPU's... it can create gaps between the slot and the contacts, as well as other issues.
A new VBIOS version would come from MSI.
 

cseck28

Distinguished
Jul 21, 2015
6
0
18,510
0
Ok, I tried moving the card to a new PCI-E slot and I still have the problem. It doesn't appear to be sagging, it seems to be firmly in position even when the tower is upright. As far as updating the VBIOS goes MSI doesn't seem to have downloads for it anywhere in their support page for my card
https://us.msi.com/Graphics-Card/Radeon-RX-480-GAMING-X-8G/support
The closest thing I've found for a VBIOS updater is this, but it doesn't seem like it's MSI's thing.
Forgive me, I've never actually updated a VBIOS for a GPU, only for motherboards. According to the AMD sofware the BIOS version I'm using is 015.050.000.000. Not sure if it's the most recent one

At any rate if it was a faulty GPU issue wouldn't it be able to recover without the entire computer crashing/freezing or I would at least be getting BSODs and crash dumps? I've been led to believe the lack of this sort of thing is indicitive of a PSU issue, which would be nice for me because they're much cheaper. Also one symptom I forgot to mention, when it happens the GPU fans kick in seemingly at max speed and faster than I ever normally see them but like I said my temps really only reach about 75C max when I'm monitoring them as it happens (which is about as high as it usually under load).
 

A_Goat

Distinguished
Ok, I tried moving the card to a new PCI-E slot and I still have the problem. It doesn't appear to be sagging, it seems to be firmly in position even when the tower is upright. As far as updating the VBIOS goes MSI doesn't seem to have downloads for it anywhere in their support page for my card
https://us.msi.com/Graphics-Card/Radeon-RX-480-GAMING-X-8G/support
The closest thing I've found for a VBIOS updater is this, but it doesn't seem like it's MSI's thing.
Forgive me, I've never actually updated a VBIOS for a GPU, only for motherboards. According to the AMD sofware the BIOS version I'm using is 015.050.000.000. Not sure if it's the most recent one

At any rate if it was a faulty GPU issue wouldn't it be able to recover without the entire computer crashing/freezing or I would at least be getting BSODs and crash dumps? I've been led to believe the lack of this sort of thing is indicitive of a PSU issue, which would be nice for me because they're much cheaper. Also one symptom I forgot to mention, when it happens the GPU fans kick in seemingly at max speed and faster than I ever normally see them but like I said my temps really only reach about 75C max when I'm monitoring them as it happens (which is about as high as it usually under load).
It's been awhile since I've used that software, but I believe once you download that software from techpowerup it will tell you what's what with the VBIOS versions.
It may hand you whatever VBIOS you need, or you may need to find it elsewhere if MSI doesn't have it. And please note that flashing VBIOS is not free of risk, I'm recommending this only because your GPU is otherwise unusable for gaming in its current state anyway.

Regarding the freezing, typically if it's a PSU issue the computer will physically power off and sometimes reboot on it's own.
If it freezes or black screens, this is typically a GPU error. Only in slight stability errors would the GPU be able to recover, such as with a blinking screen.
BSOD and dumps would typically be CPU/chipset/RAM related.
These are not always set in stone, but are common relations.

Both of these issues are hard to diagnose without replacement parts.
 

ASK THE COMMUNITY

TRENDING THREADS