What causes in-game reboots?

T

That's_Corporate

[H]ard|Gawd
Joined
Sep 30, 2006
Messages
1,738
My 4870x2 has begun acting up in Mass Effect 2 and I need some help troubleshooting.

Out of nowhere, typically with a bunch of particles on screen, the monitor will switch to black and I get a system reboot. I did some investigating and this is what I found:

CPU temps & load : All look fine. I'm Prime95 stable on my OC so I'm not too worried.

GPU temps: Never goes above 60, pretty good for a furnace. I'm on stock clocks too.

GPU load: This is where it gets interesting. In GPU-Z, the first chip seem to see 80% max usage during hectic sequences whereas the second can get pegged at 100% for quite some time. In doing some logging, that seems to be the common thread; When the second GPU is maxed for more than about 10-20 seconds, I get a reboot.

This leads me to believe it's a power issue. Then again, I have both the 6 and 8 pin hooked up to a Silverstone 750W unit. Should be plenty for a Q6600 @ 3.2 and a 4870X2.

Ideas guys?
 
Not positive, but I had a similar issue after I installed the new 10.12 drivers. I think the new CCC is doing some kind of funny conflicting.

I had to clean out everything and install the newest hotfix (as a standalone). That fixed my issues.
 
Nanan said:
Does the fan spin up to max when the screen goes black? Do you see any artifacts before the crash?

My previous post about my 4870x2 issues.

http://hardforum.com/showthread.php?t=1569907
Click to expand...

Actually no, the whole darn thing just turns off, then spins back up like a normal reboot.

PynkFloydd said:
Not positive, but I had a similar issue after I installed the new 10.12 drivers. I think the new CCC is doing some kind of funny conflicting.

I had to clean out everything and install the newest hotfix (as a standalone). That fixed my issues.
Click to expand...

Hmmm... interesting. My drivers are the older 10.11's though...
 
It sounds like a power issue. If it was overheating then you'd normally expect artifacting at times, not just silent reboots.

First thing to try is to unplug and replug the cables into the card. They can look and feel like they're seated properly, but might just possibly be making intermittent contact e.g. the contacts on the card heating up and moving away from the pins in the plug. Unlikely, but a cheap thing to try.

Do you have another PSU available that you can swap in for a while? It could be that the PSU is on the way out (even if it's very new, it's possible).

Finally, either the video card or the motherboard could be on the way out. Working out which (if either) can get very tricky - hope it's one of the other two ...
 
PynkFloydd said:
Not positive, but I had a similar issue after I installed the new 10.12 drivers. I think the new CCC is doing some kind of funny conflicting.

I had to clean out everything and install the newest hotfix (as a standalone). That fixed my issues.
Click to expand...

Well I cleaned out everything and went with the 10.12a set with no luck :(

magao said:
It sounds like a power issue. If it was overheating then you'd normally expect artifacting at times, not just silent reboots.

First thing to try is to unplug and replug the cables into the card. They can look and feel like they're seated properly, but might just possibly be making intermittent contact e.g. the contacts on the card heating up and moving away from the pins in the plug. Unlikely, but a cheap thing to try.

Do you have another PSU available that you can swap in for a while? It could be that the PSU is on the way out (even if it's very new, it's possible).

Finally, either the video card or the motherboard could be on the way out. Working out which (if either) can get very tricky - hope it's one of the other two ...
Click to expand...

No spare PSU over here so I tried to disconnect and reconnect the power cables to the card. To my absolute surprise, I was actually able to make it to some gameplay! For reference, that's MUCH farther than I could before. I was so hopeful, my heart was soaring, and then *bvvvvvvvvvvvvvvvvrrrrrwwwwww....*. If that isn't clear, my computer shut off. So still no luck :(

Looks like a new PSU for me since I'll need one for my new build anyways.
 
That's_Corporate said:
Well I cleaned out everything and went with the 10.12a set with no luck :(



No spare PSU over here so I tried to disconnect and reconnect the power cables to the card. To my absolute surprise, I was actually able to make it to some gameplay! For reference, that's MUCH farther than I could before. I was so hopeful, my heart was soaring, and then *bvvvvvvvvvvvvvvvvrrrrrwwwwww....*. If that isn't clear, my computer shut off. So still no luck :(

Looks like a new PSU for me since I'll need one for my new build anyways.
Click to expand...

That sucks... :(

Just throwing ideas out...but, have you checked for a firmware update for the card? ...and I know you mentioned the GPU load prior to reset. Just for the hell of it, have you tried underclocking the card and seeing what happens? If it's a power issue, the lower draw might prevent a reset (possibly).

I'm not sure if this would help, but Everest...err Aida64 gives very detailed system info. I can't remember if it gives the specific real-time voltage of the GPU though. If it does, it might help you get a little more info to diagnose your problem.

http://www.lavalys.com/products/everest-pc-diagnostics/
 
PynkFloydd said:
That sucks... :(

Just throwing ideas out...but, have you checked for a firmware update for the card? ...and I know you mentioned the GPU load prior to reset. Just for the hell of it, have you tried underclocking the card and seeing what happens? If it's a power issue, the lower draw might prevent a reset (possibly).

I'm not sure if this would help, but Everest...err Aida64 gives very detailed system info. I can't remember if it gives the specific real-time voltage of the GPU though. If it does, it might help you get a little more info to diagnose your problem.

http://www.lavalys.com/products/everest-pc-diagnostics/
Click to expand...

The downclocking seems to have worked! I hacked 50 mhz off both the core and the memory and I got through the most hectic scenes of ME2 without issue!

So my next question is, what does this mean?
1) The PSU is messed, rail perhaps?
2) The card is toast.
 
first take a picture of your wiring of your comp.

Than I will either make fun of you or say it is a bad power supply.
 
Whenever I had this problem, my CPU overclock was unstable. Prime95 stable does not mean game stable. Try increasing your vcore; that always works for me when I have this problem.
 
Sounds like you have a bad core on that card. 50mhz decrease would barely drop power usage unless there was a significant decrease in voltage too.

Honestly though, I don't know how you get 60C under load. I know it is doable, but that is great on an X2.
 
TheBuzzer said:
first take a picture of your wiring of your comp.

Than I will either make fun of you or say it is a bad power supply.
Click to expand...

Hahaha! That USED to be how I roll, but then I discovered the wire management cult on [H] and now my case is very clean. Everything is tucked behind the mobo tray.

So, bad power supply it is :)

jbz7890 said:
Whenever I had this problem, my CPU overclock was unstable. Prime95 stable does not mean game stable. Try increasing your vcore; that always works for me when I have this problem.
Click to expand...

Well CPU-Z is showing my core voltage as 1.336. I swear I set it up as 1.35 in the BIOS though. My overclock isn't anything major though, just 3.2 ghz, and that's on a G0 revision.

I should note that this reboot problem was around BEFORE I noticed that my BIOS settings were lost somehow, and I was gaming on all AUTO settings for my CPU (stock 2.4 ghz, with auto voltage).

jamesrb said:
Sounds like you have a bad core on that card. 50mhz decrease would barely drop power usage unless there was a significant decrease in voltage too.

Honestly though, I don't know how you get 60C under load. I know it is doable, but that is great on an X2.
Click to expand...

Yeah... that's what I was thinking too... Such a low GPU drop doesn't sound like it should work in theory...

Oh, and 60C under load is easy, it's called 70% fan speed haha
It's okay though, like Tali in Mass Effect 2, I was born on a space ship, so I'm used to loud noises constantly.
 
Will just be vdroop, 0.014V drop is't much to worry about, unless of course you're running the volts really fine for your overclock. I'll put it out there that I used more than that for my Q6600 at that speed, 1.45V in fact, but it was a bad stepping.
70% fan speed is pretty epic on a 4870X2 but, I have to say, latterly my cards ran higher than that (100% in fact) on auto in more demanding games like L4D2, this is of course in crossfire, but with huge amounts of case ventilation (6 1900rpm Scythe slipstreams plus a 140 and 230 coolermaster). I'd keep an eye on your temps in the more demanding games, just in case they aren't much higher than you think they are.
 
Well I'd be willing to try more vcore if that might help.

How much is safe on a Q6600 with a TRUE attached?

Oh, and the rebooting still occurred at 100% fan speed on the 4870X2. At that speed the temps never got above 45 degrees I believe.
 
if your system is rebooting during gaming (heavier power draw), time for a new PSU.
 
The system I just built was experiencing the same issue. It was the new power supply. Rma'd it and now everything is great.
 
It can be the PSU, but not always. Usually I find if it's the PSU, the system shuts down and then reboots, rather than just instantly rebooting.
 
That's_Corporate said:
The downclocking seems to have worked! I hacked 50 mhz off both the core and the memory and I got through the most hectic scenes of ME2 without issue!

So my next question is, what does this mean?
1) The PSU is messed, rail perhaps?
2) The card is toast.
Click to expand...

Good to hear!!!

I'd second what the others have to say about your CPU overlock. Try tweaking your settings in general though.

There's 2 things to keep in mind besides just your PSU and GPU hardware.

1.) ATI's implementation of Powerplay (or whatever their power saving tech is called) sucks. I still have HDMI sound issues, although they're not as bad now. (...but it took like 6 months after the 58xx's release to get working drivers!) Play with your settings in the Overdrive section of your GPU drivers and hopefully you find a sweet spot.

2.) The card does get some power from the PCI slot. Maybe that dual-GPU card you got is starving? I think my motherboard allows me to bump up the voltage on my PCI slot. Maybe do a VERY small bump on the voltage and see if that fixes it?
 
YeuEmMaiMai said:
if your system is rebooting during gaming (heavier power draw), time for a new PSU.
Click to expand...

...and it's common for PSU's to start decaying after about 2 and a half years? I got the Silverstone because I thought it would last.

tikiman2012 said:
The system I just built was experiencing the same issue. It was the new power supply. Rma'd it and now everything is great.
Click to expand...

Strange thing is that I've been using this PSU for a while. I guess there goes the whole RMA idea hehe

samuelmorris said:
It can be the PSU, but not always. Usually I find if it's the PSU, the system shuts down and then reboots, rather than just instantly rebooting.
Click to expand...

Mind explaining further? I mean... everything sounds to switch off, then turn back on. Not sure how that's different than an 'instant reboot'.

PynkFloydd said:
Good to hear!!!

I'd second what the others have to say about your CPU overlock. Try tweaking your settings in general though.

There's 2 things to keep in mind besides just your PSU and GPU hardware.

1.) ATI's implementation of Powerplay (or whatever their power saving tech is called) sucks. I still have HDMI sound issues, although they're not as bad now. (...but it took like 6 months after the 58xx's release to get working drivers!) Play with your settings in the Overdrive section of your GPU drivers and hopefully you find a sweet spot.

2.) The card does get some power from the PCI slot. Maybe that dual-GPU card you got is starving? I think my motherboard allows me to bump up the voltage on my PCI slot. Maybe do a VERY small bump on the voltage and see if that fixes it?
Click to expand...

No way, you can tweak the ATI power saving stuff?! Where do I do that? Can't seem to find it anywhere in the Catalyst suite.

I'll try the voltage bump. But it makes me wonder why this is becoming an issue now. If it was a voltage tweak issue, wouldn't that have showed up before when I built the system back in late '08?
 
Mind explaining further? I mean... everything sounds to switch off, then turn back on. Not sure how that's different than an 'instant reboot'.
Click to expand...
Yeah that's exactly what my crappy BeQuiet did. Entire system powers off, including fans/lights, then 2-3 seconds later powers back on.
What I mean by an instant reboot is one where the graphics fan speeds rise like they do when you first power the system on, but the system remains on, just with the screen blank, then you see the BIOS screen, and it boots back into windows.

I should add here, I also experienced similar behaviour with overheating issues when I had a dustbunny in one of my X2s, however on one occasion. the system did not actually power off, it just rebooted. On the other occasion it did power off.
 
Last edited:
That's_Corporate said:
No way, you can tweak the ATI power saving stuff?! Where do I do that? Can't seem to find it anywhere in the Catalyst suite.

I'll try the voltage bump. But it makes me wonder why this is becoming an issue now. If it was a voltage tweak issue, wouldn't that have showed up before when I built the system back in late '08?
Click to expand...

I meant your fan/overlock. The one thing that I remembered from past driver updates is that they've been tweaking Powerplay (not sure if it's also been for the 48xx series) over the last bunch of Catalyst packages. I remember a few which I had to play with the slider and then turn it back to where I wanted it. For some reason, that reset it to defaults. It was some weird bug with CCC.

There is a way to edit the profile manually to override settings, but I don't see much benefit to that. You can also try out the GPU clock tool since that ignores Powerplay, if you're interested.

Yeah, not sure why it started acting up recently? If all else fails, you can always backup everything and try a clean install of Windows if you still can't narrow it down.
 
Okay well I hate to resurrect a dead thread, but I have some new information.

At the suggestion of a fellow [H] member, I ran furmark and LinX at the same time, to simulate a maxed GPU and CPU condition while gaming. To my complete shock, everything went off without a hitch for the full 10 minute test, which is both longer than how long I can last in-game with Mass Effect 2 and at much much higher loads on both the CPU and GPU according to my logging.

To me, that seems to rule out both the overclock, bad 4870x2 card, and PSU, don't you think?

...or are in-game conditions somehow different than GPU and CPU load testing software? Perhaps less demanding on the PSU even though both the GPU and CPU loads are maxed?

I'm completely stumped :p
 
It sounds like overheating, but since it's not your GPU or memory, it probably is your VDDC Phase temps. I had problems with these on my old 4870 CF setup. Sadly there is nothing you can really do sans maybe changing the padding.
 
JMCB said:
It sounds like overheating, but since it's not your GPU or memory, it probably is your VDDC Phase temps. I had problems with these on my old 4870 CF setup. Sadly there is nothing you can really do sans maybe changing the padding.
Click to expand...

Hmmm, what's the acceptable range for VDDC Phase temps?
 
Run Furmark for longer than 10min, about 30 minutes. Try different games out and see if the same problem happens. I think you said this happened when you upgraded drivers so reverting back to the old ones might help. I don't agree it's always a PSU problem especially if you are running furmark for an extended amount of time. Exhaust all your options until there is only one option of getting a power supply.
 
Do you get a Windows Recovery message when you restart? If so what does it say?

Anyway, I still think it's a CPU/RAM overclocking issue. Perhaps you're running your RAM at too high a speed. Try setting everything on your computer to stock speeds and then run some stress tests and play games. If you still experience crashes, then we would be able to rule out overclocking.
Posted via Mobile Device
 
ravensteel said:
Run Furmark for longer than 10min, about 30 minutes. Try different games out and see if the same problem happens. I think you said this happened when you upgraded drivers so reverting back to the old ones might help. I don't agree it's always a PSU problem especially if you are running furmark for an extended amount of time. Exhaust all your options until there is only one option of getting a power supply.
Click to expand...

I think the driver issue was it!

I previously had this issue on 10.11 and 11.1, so a fellow [H] suggested I go all the way back to 10.9 and so far it's been completely stable. Only issue is: Performance in ME2 and borderlands is now HORRIBLE!

Any way to solve this? I think I may need to ditch the 4xxx series... new ATI drivers seem to hate them.
 
You must log in or register to reply here.
Back
Top