Alright, I'm at my wit's end with this crap. Whenever playing video games, my PC just shuts down. No errors, no warning. Just off and monitor black. To memory, this happens with Half-Life 2, Champions Online, Portal, and most recently and most often, Aion.
I've ruled out cooling issues. I have two internal fans, not including the processor fan and the power supply fan. Besides that, out of sheer desperation, I've removed the side of the case and pointed two fans directly inside the computer.
I ran MemTest for about an hour and found no issues with RAM.
Posts
Use Coretemp to monitor your heat and see what's up.
XT2700 I think.
it never did it when It was just idling on the desktop or browsing the web, but anytime I played a 3d game after around 10 minutes the machine would just turn itself off.
Replacing the card fixed the problem
naknaknaknaknak
It could also be the fan/core of your video card, typically since there is so much power being fed through those guys, if that guy starts to overheat, it can trip the safety-shutdown trigger in your power supply. Typically though you'll start to get artifacting real bad in the games before that happens....
Are you getting any weird graphical glitches in those games before the shutdown hits?
Wud yoo laek to lern aboot meatz? Look here!
Perhaps it is cooling. Right now Coretemp is showing 90 degrees farenheit, which seems like a lot considering I'm basically violating my PC with fans.
Nay on the graphical glitches. Everything runs perfectly... insanely so.. and then... nothing.
90 degrees farenheit is pretty cool. Thats 32 celsius, which is no worse than a hot summer day up here in canuckistan.
Edit for clarification - depending on processor, the max temp threshold is up at 70-90 celsius, so more than double what you're getting unless coretemp is reporting in celsius.
Alternatively, it could be your power supply on the fritz.
Certainly:
AMD PHenom II Quad Core Processor
EVGA Geforce GTX 260 Video Card
Gigabyte GA-MA790X-UD4P motherboard
4 GB DDR2 RAM
Antec EA650 PSU
http://pcsupport.about.com/od/toolsofthetrade/f/powersupplytest.htm
Takes a multimeter or special power supply tester to do.
1 Western Digital SATA Harddrive
Onboard sound
1 LG 22X DVD+R
No external drives
One fan on top, one fan in the back. One CPU heatsink (I checked, it's functioning), one power supply fan. One or two fans situated around the case, opened, in various attempts to supply more cooling.
One power light on the front.
EDIT: I'm preparing to rule out cooling, leaving it between the graphics card and the power supply. Here are some Core Temp logs, listed in Fahrenheit. In both instances, Core Temp is started right before booting Aion, and ends right before my computer shuts down:
It's happening much faster now. In Portal, I only crashed the one time. Granted, it was a short game.
In Lord of the Rings Online, I'd crash once every few days.
In Champions Online, I crashed about once every other day.
Earlier today with Aion, I woudl usually get at least 15 minutes, sometimes even half an hour. Now it's nearly instantaneous.
If you ever need to talk to someone, feel free to message me. Yes, that includes you.
Power Test
I'll check out that link tonight, Red Rover. It's blocked at work for some reason.
I guess I need to buy a new graphics card?
edit: maybe I'm just not as well versed in the most modern CPUs as I used to be, but under load 105F seems suspiciously low to me. Are the modern quad cores that cool? I only say that because the only issues I've ever encountered with PSU's or sometimes bad memory modules is that they will simply reboot the machine rather than shut off. Then again it's been about 3 years since Ive really been in the game when it comes to troubleshooting hardware. Also with your temp logs, you're playing a game for over 40 minutes and the core temp only increases 1-5 degrees F? That seems even more suspicious to me. What it sounds like to me is you had a core meltdown (happens much more often now than it did in the past.) A quick cursory google search yields that even hardcore overclockers with the Phenom II 940 and the best air and water cooling are idling higher than your under load which leads me to doubt that those temp readings are in F.
Holy crap you're right I just noticed that. My bios also shows in Celsius first and if those temps are in C no wonder it's shutting down faster and faster. The core is still hot from the previous session so it's overheating faster and faster.
Prof can you see if those temps are being shown in F or in C? If those are in C then you're dealing with a very serious CPU core overheating issue, it could be something as simple as just your CPU cooler not seated properly or it could be something much worse. We'll need to know a bit more to be sure.
Wud yoo laek to lern aboot meatz? Look here!
That being said, there are certain intricacies concerning reading the temperature of a Phenom quad-core that I may be failing to grasp.
This may be a red herring, but tonight I experienced no issues after 2 hours of playing Aion. After removing the power supply and testing it, I noticed a switch connected to each fan, that allowed me to dictate the power of the fans. I set them all to high.
If that's what is making the difference, then it was cooling all along and I've been approaching this the wrong way the whole time.
But like I said, it may be a red herring.
EDIT: Bah, Core Temp is still showing around 91 degrees Farhenheit. I don't fucking know.
Huh...well that's a stumper then, as others have said, that definitely sounds REALLY really cool for a cpu under load, farenheit wise, most cpus reach 50-70 degrees celsius or 120+ degrees farenheit. I'm really not sure anymore beyond, trying to swap out a PSU or video card and see what happens.
Wud yoo laek to lern aboot meatz? Look here!
If that's what fixed it, and I use that term loosely, you're still in the danger zone. Running the case fans on a higher voltage will typically only lower the temperatures by a couple degrees C.
Sorry if I stick to my guns that this is a CPU issue, but there are so many red flags here. Try running the most intense 3D game you have (Crysis or etc maybe?) rather than MMOs and post the resulting temp logs. Most MMOs aren't nearly as taxing as stand alone games. Are you using the stock cooling unit on the CPU with the stock heatpad?
Also try using SpeedFan instead of CoreTemp. Then post the logs with not only temperatures but voltages as well. If voltage spikes are causing this hopefully we can see that.
Also, if your cooling is overkill for your system, the difference between idle and load can be quite small. My system idles at 36 and hits 40 maybe at load because of the way I have it set up.
Those MMOs are the most graphically intensive games I own. If you're referring to the heatsink, I am using the one that came with the processor, yes.
I'll give this a shot when I get home tonight.
1: Really really bad software fuckup. Very very rare. Also possible from RAM/CPU deaths, but usually that bluescreens you instead.
2: Heat. Fairly common, a fan breaks, etc. Sounds like you've eliminated that options.
3: Power. Sounds likely given your heat info. What's the specs on your power supply unit?
You can find the full info in the URL I provided.
We don't have a lot to go by here to help you unless you can start testing and ruling some things out.
You make it sound like I haven't been testing and have been resistant to advice. I've used two (now three) temperature-monitoring programs. I've used a power supply tester.
3DMark apparently does not work with sub-Vista operating systems, so I settled for another Aion test, this time while running Speedfan.
Here's about an hour's worth of SpeedFan log. Temperature values are in Celsius.
I wouldn't study it too closely. Just about any point in the middle will be an indicator of the whole. The voltage values read like greek to me, but people were saying there may be something to them, so there they are.
The temperature ranges are pretty consistent with what I was seeing in Core Temp before my dubious "fix."
I can't justify purchasing a new graphics card if the problem doesn't return. I'm aware this is highly nebulous, but if it's working, I just can't lay out that kind of money. I'll keep on watching and testing in the meantime.
EDIT: I'll run Prime95 overnight and during work and see if anything comes up.
Absent actually switching out hardware, there's not a whole lot left to troubleshoot through software means. Though I would still be interested in seeing some logs from Prime95 (which will push your machine to the absolute limits of its capacity which almost no game can do.) Otherwise if you can't recreate the problem, just keep an eye on it and come back here if it starts acting up again. Best of luck.
Nothing of note, I'm afraid. I built it a couple months ago and it's been a great machine, save for the recent hard shutdowns.
Sorry for getting defensive, Simpsonia.
You should be able to find a place that'll let you return them if they don't fix your problem, but I'd bet it's one of the 2.
Since I can't seem to replicate the problem, despite it's relatively long history, I wouldn't be able to confirm if either of those solutions help or not.
It's been three days where I've had hours of gametime without any crashes. Stress testing isn't producing results and cooling continues to appear fine.
I hate to leave these things hanging without a clue, but the problem has simply ceased. If it repeats, I'll look into replacing the power supply or graphics card. In the meantime, however, it looks like it's Resolved (?).
I want to thank you all for your help and patience. Even though the actual solution is mysterious, I learned a lot in the process. You guys were very helpful. Thank you so much.