Random reboots. GPU or PSU?

S

spaceman

[H]F Junkie
Joined
Jan 7, 2005
Messages
14,925
Recently went from a XFX 290 DD to a Zotac 980 AMP!
http://www.newegg.com/Product/Product.aspx?Item=N82E16814150697
http://www.newegg.com/Product/Product.aspx?Item=N82E16814500361

Psu is a XFX Pro 850

http://www.newegg.com/Product/Product.aspx?Item=N82E16817207028

I have selected high perf in windows, tried each oc tool to force overvoltage and also run the card with no oc and no overclocking software either.

PC still reboots. I have not seen or been able to find a crash report.

The pc just reboots. No bsod. The only other weird thing it did was have sound issues. Like you could hear the gpu running or just power flow effected static.

In game, no crashes. Just in either windows idle or when I load the start screen to a game.

Derp? I thought it was some setting. It appears to be something else? The two gpus use different power plugs. The 290 used two more actually.
 
The 290 pulls more juice for sure, but since you moved to different connectors on your PS with the 980, perhaps it is power related? Have you tried using/swapping any of the other power connectors on the PS for the card? Perhaps you have a rail that isn't quite holding up all that well and by using different plugs you switched to that weaker rail? Try using some of the other ports or the same one that the 290 used. How old is the supply?
 
Trying the connectors. Just made the first switch and it rebooted. Trying the other now. PSU is two years old.
 
Thats way more than just an incorrect setting...especially being random. If it only happened under certain type of loads/stress then it would give us more to work on. Use an inexpensive multi meter and check the rails on the power leads...if i had to guess you might find the rails are way out of wack. We need to know what the 12v and 5v rails are doing idle and while stressed....if there measuring 4v or 11v or outside of tolerance we can pretty much say the supply is gone bad. If it were the gpu, i think it would happen under 3d stuff like in games and benchmarks only.

If no meter is available at least check and see what software is saying about the rails...gpu-z has a sensor tab that should be looked at.
 
Last edited:
lol

I think the soundblaster zx was loose and touching the back of the gpu. I got the same sound issue after swapping psu connects and saw it was leaning way over. Pretty dark in my man cave. My case is a test bench so yeah. Great airflow but kind of dark and cramped.

Derp. I think that got it.

Nope. Still did it even after that. I did not have the oc or power overvoltage enabled.

Rebooted and loaded the OC and just played a round of mechwarrior with flawless performance. Weird.
 
Last edited:
Going to buy a new psu. Might not be the issue but it sure seems likely. About to get snowed in again so no way I want to be without my gaming rig.
 
Yeah. No PSU for me. My city has one pc shop worth a shit and it is closed. Great.
 
Just reinstalled Win 10 as this appears to be a new "feature". Probably corruption or some other wonderfully coded update.

yay
 
Put your old card back in. Does it stop rebooting? If so then it likely isn't the power supply.

Are you overclocking your CPU? If so, set that back to stock and see if it crashes first.
 
Last edited:
Got a Corsair 750 and it is running fine so far. Shrug. Guess it was the psu.
 
Just did it again. Not even under load. Just getting teamspeak open. Crap. Trying a driver re-install.
 
spaceman said:
Just did it again. Not even under load. Just getting teamspeak open. Crap. Trying a driver re-install.
Click to expand...

RIP GTX 980 Amp!
2016-2016 :(

Ok but seriously since you mentioned you could here static from the sound card... Remove the sound card or move it to a different PCIE slot. Could be a grounding issue or something odd.
 
Yeah. Just took the ZX out. Loading realtek now. Going to be PISSED if this is a Win 10 issue. Might be a mobo bios compat issue. Running an old MSI P67A-GD65 mobo.
 
Creative Labs screws up everything! :eek:

I had a similar MSI P67A G45 board and almost always had static through my Sound Blaster Z (which is now sitting in a box) when I was trying to use a GTX 970. Didn't matter which PCIE slot it was in or what drivers I was using. Using that motherboard I didn't have the issue with the GTX 660 or multiple AMD cards I was using before a 970. I gave up on the Z and eventually switched to a newer platform. The new Realtek 1150 hardware is good enough I haven't bothered trying the Z again on my Skylake build. However this may need more time to draw conclusions, you could find a solution or the issue is unrelated.
 
Last edited:
I always hated Creative Labs drivers, some of my worst memories with troubleshooting led to them. But if its a hardware issue I'd put my money on the 980.
 
lol is there really a huge difference between a z or zx card driver? the driver has minor issues but overall its not bad with 10.....he got something else going on i think
 
It appears to be the gpu or my mobo being too damn old. Last bios update was in 01/2013. Nothing new since then. Still having the same issue. Just random. When games play they do so perfectly.
 
Last edited:
Do you use DDU to uninstall the old AMD/Nvidia drivers? Great tool to use in Safe Mode to clean out old video driver sets.

But seriously, if you swap to the old 290 and it works perfectly fine then it is the 980 having issues. Your motherboard is a P67 SandyBridge / IvyBridge board - it shouldn't have any issues running a 900 series Nvidia card (my old X79 Dark doesn't, and it also runs SandyBridge and IvyBridge based CPUs).

You COULD try verifying that the PCIE setting in BIOS is set to GEN2 if you have a SandyBridge CPU, or GEN3 for IvyBridge, instead of AUTO. Strange as it sounds, some folks have had issues with the 900 series cards and the AUTO PCIE setting.
 
I sold the 290. I have an 0ld 240gts in my media rig upstairs but that is pci.

Weird. There is no pci-e in my bios to select?

MSI P67A-GD65

Just played for about an hour. No issues. Great performance in Tomb Raider. Then reboot in windows. Then again in windows. No error messages. Very frustrating.
 
Might be ram related. Disabled extreme mem timings in bios and it has been stable so far. I did not have the cpu oc ed at all. The gpu is stock now too.
 
If you were overclocking the GPU, it can lead to issues transitioning between power states and reboot the PC. One thing it keep in mind is that the card is almost always switching power states based on the load, even within the same game. This can happen at high or low GPU loads without much of a warning. This comes up a lot when people are pushing cards close to 1500, the card can be 100% stable in Furmark and then simply crash opening a browser window or playing a game. There is a huge community of people doing BIOS mods over on www.overclock.net to avoid these issues when overclocking Maxwell GPUs.
 
It was happening at stock clocks except that the ram had extreme timings selected in bios. I disabled that and it has been stable ever since.
 
spaceman said:
It was happening at stock clocks except that the ram had extreme timings selected in bios. I disabled that and it has been stable ever since.
Click to expand...

It's been barely 24 hours since this started, wait a couple days before declaring it stable :p
 
No issues still. That must have been it. Weird. Time for new ram I guess. Probably time to update the mainboard and cpu too really. 2500K has been with me since 2011 lol.
 
I also suspect ram being somehow related to whats going on...but this can be troubleshooted without totally replacing all the ram. First thing i would try is under clocking the ram.....relaxing the timings, switch to or from xmp in bios. The issues your describing is verbatim of what happens on mine when i overclock my ram past a certain point - windows reboots out of know where...almost exactly like a bad power supply. When you changed to nvidia card its now using ram differently, maybe even parts of ram that the amd card never got to. you can also try removing half your ram and see what happens...might be just one bad stick. Ocellaris is correct...its takes a full week to really declare the pc fixed
 
Yeah. I need to upgrade to 16 gb anyway. I will buy some and test the old ones to see which one is naughty. Still running perfectly btw. XMP is off. All oc's are off.
 
Passed the test! I guess the ram just can't hold the tighter timings anymore. At stock it is fine apparently.
 
Just used the Intel OC software to boost things back to 4.2 and EVGA Precision to tune the amp back to 1400 core clocks. Runs Tomb Raider fine.
 
well well well.....what do ya know. Is was the memory settings after all ;) I thought it was very possible
 
Yeah weird. Especially when the timings have been the same for 5 years. It must be like you said.
 
Since I sold you the Zotac 980, it's good to hear the GPU wasn't the problem. :) Glad you got it fixed.
 
Had a near identical issue recently caused by; wrong oc settings after bios flash, a gt610 fighting/coexisting with a 290x and a loose molex>sata adaptor. Tackling all three got me back rock solid. Went from hard reboots to soft lock/screen freeze sound still work, then finally stable for days on end like it should usually.
 
You must log in or register to reply here.
Back
Top