A game causes my PC to shut off and can only start after clearing CMOS

[Nazralte]

So, I installed Assassins Creed Unity and it ran fine a couple of times. It crashed once but started back up fine. A day later I go to play and right before the menu (press a button to start) screen comes up, the PC turns off and will not start. A friend of mine suggested I try clearing the CMOS first before assuming it's the PSU. I do that and the PC starts right up.

So I think that maybe there is something odd in the install. I reinstall the game and again, it plays a few times and a day later, PC turns off and I have to clear the CMOS to boot it back up.

I know it's not a heat issue. I have ran both GTA V and Skyrim (heavy mods) for hours and this has never happened.
What could be causing this? I have never had this happen in my 20 years of gaming.

I have a Gigabyte GA-970A-UD3 (rev. 1.0) (bios is updated)
An AMD Phenom Black 970
Nvidia GTX 970
550W PSU
8gb of ram (two 4gb sticks)
Samsung SSD
WD Caviar 1tb (this is what i run my games off of)

 

[Sywtch]

Without knowing the exact scenario i have to say the weak link here is that PSU. What brand is it? How old is it? With your board/Chip pulling 300+ watts (http://www.techradar.com/reviews/pc-mac/pc-components/motherboards/gigabyte-970a-ud3-978161/review/2), and a video card that wants 150ish..add in spinning discs, a few fans and other accessories and you are looking at hitting right up against the top end of what the PSU can supply. Typically when you are planning a PSU out for a new build you follow the 75/25 rule. Get a PSU that can serve your desired system wattage under load and have 25% of the PSU total wattage left over.

I would suggest hooking something like a kill-a-watt meter up to your outlet to get your acutal wattage while gaming and see how close to 550 you are.

 

[Nazralte]

Well, again, this is happening at the menu screen, before anything stressful happens. And again, I run Skyrim, heavily modded and GTA V without any issues. Both of those stress the system. I hear the fans on the video card more than any other game I've played due to all the mods I have to make the game look better.
Also, I used a couple of PSU calculators online (Cooler Master and http://powersupplycalculator.net/) and both show about 400 - 450. Mine is a ZuMax 550 and it is a few years old. Thing is, if it was a PSU problem, it would happen with the other games, so I sort of have to rule that out. I can't believe that running those other games don't cause it but the menu screen of AC:U causes it the day after I played the game for 2-3 hours fine. Doesn't make sense.

 

[Sywtch]

Its possible, I am just pulling from one data point to be honest. I have had this happen to me before and it went away with a new PSU. The only other thing i can think of is some kind of thermal issue. The not powering up until you clear the CMOS could be indicative of some kind of surge or voltage protection on the motherboard side. I was reading the other day about surge protection built into some new ASUS boards these days, does your motherboard manual mention anything about a feature like that?

I would say that from what you are telling me everything screams hardware issue. I know its only one game, but i just cannot fathom what could possibly be done in software to cause the BIOS/CMOS to need to cleared. That bit just doesn't make sense.. Do you have any of the motherboard utilities installed that might be able to interact with the BIOS (some of the tweaking/oc software perhaps?).

Other than that i would start trying to narrow down component issues, do you have any spare parts you can use to test? A spare set of ram? Maybe try running via onboard video or an older video card for a while?

 

[Nazralte]

I still can't think it's hardware at the moment. Because this is happening before the game loads anything at all. Nothing is being stressed.
I think the next step (unless I just avoid playing the game for now) is to check my motherboard drivers. Or perhaps there is some error log being generated I can look at and see if there is a driver fault.

 

[westom]

A day later I go to play and right before the menu (press a button to start) screen comes up, the PC turns off and will not start. A friend of mine suggested I try clearing the CMOS first before assuming it's the PSU. I do that and the PC starts right up.

Nothing in the install is even looked at until the power controller powers up hardware, establishes hardware integrity, and finally lets the CPU execute. Your power system (which is more than just a PSU) is not even gettting that far.

First suspect is a CMOS battery. A meter provides a best solution here. Measure that CMOS battery (ie a coin cell). A voltage above 2.9 means it is good but aging.. A voltage slightly lower means battery is failing - will need replacement in months. An even lower voltage (maybe 2.7 or lower) means a defect has been identified - without speculation or doubt.

Second, if something is wrong with the power 'system', then that meter and some requested instructions will identify which 'system' component is defective without doubt. Why it is defective. And fix it the first time.

No heat issues, bad software install, defective disk drive, or many other suspects that get accused only on speculation exist. You have described a defect that never even lets the CPU execute setup in the BIOS - long before any software or OS gets executed. Likely suspects are limited only to what is described above. Described is how to identify a suspect without further speculation.

 

[Nazralte]

Westom, just know, I am not being snarky or anything, I am just trying to clarify.
So you are saying that a bad battery would cause a PC to shut off at a specific point of a specific game booting up? This is a repeatable problem. It happened twice in the same exact way. Installed and play first day, second say (left PC running mind you) it shuts off and I have to clear CMOS to boot.
In my thinking, the game is somehow causing some sort of fault which is corrupting the CMOS. As to how, I have no clue, which is why I am posting here.
I don't know if it is related, but I did check the windows event viewer and for a long time now, at random times, even when I know I wasn't using the PC, I see error 500 and 501 which says
"The Desktop Window Manager is experiencing heavy resource contention.
Reason : Graphics subsystem resources are over-utilized.
Diagnosis : A sharp degradation in Desktop Window Manager responsiveness was observed."

Again, I haven't had any other problems with any other games or applications, but if there is something causing this error and for some reason AC:U interacts with it a day after being installed....

 

[westom]

So you are saying that a bad battery would cause a PC to shut off at a specific point of a specific game booting up? This is a repeatable problem. It happened twice in the same exact way. Installed and play first day, second say (left PC running mind you) it shuts off and I have to clear CMOS to boot.

We should probably go back to a subjective word - "crash". Exactly what was a crash? A crash might be a BSOD or the PC powering off (maybe then rebooting). Did you confuse "crash" with other problems. For example, some web sites today connect your computer to maybe 60 other websites. Others do not. A computer is not slow or old as so many techs assume. That computer is trying to download and send information (about you) to those other web sites including akamia, amazonaws, facebook cloudnet, synacor, yahoo, etc - all sites you otherwise never visit. These other connections mean more video and network resources are consumed. That computer appears slow. So naive computer techs insist you need a faster machine. Cut off those other vampire connections and a 1 Ghz machine is fully fast again.

Worse, some vampire connections remain existant even after you leave the site.

Did you assume hardware problems rather than software congestion? Again, a point I constantly make - follow the evidence. A computer is slow, so my power system must be defective - is a common reframe among so many computer techs who never learned how to 'follow the evidence'. Who only learned to suspect their first wild speculation. And then resort to a worst diagnostic procedure - shotgunning.

Typical UPS connects your computer directly to AC mains when not in battery backup mode. That is when power is cleanest. Some UPSes are so simply designed that even noise may cause its voltage detection circuits to assume a low voltage and switch to battery. A battery more than 3 years old may no longer work. This test is simple. Pull a UPS power cord from a wall receptacle. When it switches to battery, can it still output enough power to maintain computer and other attached appliances (ie monitor)? If not, battery is probably more than 3 years old.

Some games consume more power than others. A meter could detect that a power system was always defective - maybe when first purchased. But that defect does not appear until that more 'power hungry' game runs. Normal is for defective hardware to still boot and run a computer. Only numbers can identify the defect.

Power system is the foundation of a computer. If that 'system' is defective, then computer may very intermittently crash. That says nothing about what is on AC mains since the PSU must even provide full power when incandescent bulbs dim to 40% intensity. Power 'system' must be exonerated by a meter long before anything else can be suspect. And again, explicitly what does the word 'crash' mean?

Meanwhile, CMOS can change if its battery is marginal. Noted earlier, nothing other than a meter (the numbers) can identify that as a defect or irreelvant. Your CMOS sympotims are classic of a battery at maybe 2.7 volts. Nobody can say more without that number.

 

[Nazralte]

Westom, I never said crash, I said turn off. Which means off. There is no crash, it literally turns off like I pulled the power cord. The reason I asked my question, which you quoted, is that I have always been under the assumption that the CMOS battery was only used when the system was off to save the BIOS settings. If there is no battery (or the battery is dead) then the system would still boot but would have the wrong date, any BIOS changes would be default, ect.
And just to be clear, I can reboot the PC, leave it off for awhile, and I am sure I can even have it off without it being plugged in and it would still boot fine.

That all being said, I do not rule out the possibility that for some strange and weird reason, Assassins Creed Unity causes some odd power spike which flips out my motherboard because it hits whatever limit my PSU has. But to say the CMOS battery is to blame just seems rather odd.

 

[westom]

The reason I asked my question, which you quoted, is that I have always been under the assumption that the CMOS battery was only used when the system was off to save the BIOS settings. If there is no battery (or the battery is dead) then the system would still boot but would have the wrong date, any BIOS changes would be default, ect. ...
But to say the CMOS battery is to blame just seems rather odd.

Original post said, "It crashed once but started back up fine."

If the battery was removed, CMOS would return to a default setting. If battery was sufficient, CMOS would save current settings. But if the battery is at some intermediate voltage (ie 2.7 volts) then CMOS could be corrupted. Then CMOS must be reset to permit a startup. Corrupt CMOS parameters would make no sense to appropriate computer chips that need that information.

To say anything useful means voltage numbers from that battery. Symptoms are consistent with a battery that is too low to properly save settings and too high to return to default settings. Battery is only one possible suspect. But also explains why CMOS needed resetting. Irrelevant is whether it is or is not plugged in.

Odd power spikes are extremely rare - maybe as frequently as once every seven years. And further made irrelevant by a good CMOS battery and other existing computer circuits. Spike is often blamed because other and more common reasons are not known. Best answer (and fastest solution) comes from measuring that battery voltage.

 

[Nazralte]

Ah right. Well, the crash I referred to isn't the actual problem. And looking at it, I probably shouldn't have included it. The crash was just a simple crash to desktop, which can happen in any game these days.
The shut off and CMOS clear to power up is the main problem.
But hey, if you think that an off battery can cause a specific game at a specific time (when the game is loading the menu screen) to shut off the PC then I can always try a new battery. Cheap check.

 

[westom]

The shut off and CMOS clear to power up is the main problem..

CMOS and crash can be two separate problems. Or may share a common defect.

In a pre-emptive multitasking system, no program should cause a power off. Task (that game program) can crash. But the OS must keep working normally. However something completely different (ie a power system defect) may explain a power off or reboot. No program should cause an OS software crash or power off in a pre-emptive multi-tasking system.

That may or may not be related to the CMOS problem. CMOS may be a completely unrelated to that crash. These are curious anomalies. To say more means numbers.