My computer has been hanging for no reasons, usually when playing graphically intensive games, like Mass Effect 2 and Lord of the Rings Online but occasionally while in Windows Explorer or even when it just starts up.
Outside of games,the symptom of the problem is that the screen freezes; the display is still there, but the mouse and keyboard will not respond. It freezes less frequently while in a window application environment, but it could happen any-time. In middle of while starting Windows, showing the Windows display, while surfing the web, using Word, coding in Eclipse. It happens about once per week.
Inside games, the screen will just go black. It happens every time while playing high performance games, and after reboot, the problem doesn’t happen again but recently it will just keep crashing the game.
The CPU light remains on for both cases.
Here are some tests I have ran:
RAM: Windows Memory Diagnostic Tool and MemTest both reports error in memory once, but I haven’t been able to duplicate the results. I ran MemTest for about an hour (with 4 processes), achieving coverage of about 200% for each process.
Graphics Card: RivaTuner reports that the card’s temperature is within acceptable range.
Questions:
Are RAM errors supposed to be duplicated with the same results when using the same testing program?
Is it likely that my PSU’s output is not consistent, causing CPU operations to ‘skip’ or otherwise writing erroneous error to the RAM?
What’s the best course of action for me now? I can’t verify that the RAM has problems, as the tests’ results cannot be duplicated. What other tests should I run?
Faulty memory can be intermittent. You might not necessarily get repeatable results. RAM errors aren’t always RAM errors though. They can be CPU errors.
A bad PSU is possible. It is however only one of many possibilities.
What is your CPU temp? What is the graphics card temp? (actual temperatures, not just reported as “within acceptable range”) What’s the temp inside the case? If you don’t have utilities to get the CPU and case temp, you can usually get them by rebooting the machine and going into the BIOS setup.
A bad CPU is probably more likely than bad RAM. If you have two sticks of RAM you can pull one out and see if you still have a problem. If you do, you can pull that stick out and put the other one in. If you still have a problem then, then it probably isn’t the RAM. It’s not likely (but not impossible) for both sticks to go bad at the same time.
Another thing to try is just reseating everything. Pull out the CPU and put it back in (this will likely require you to put more thermal paste on the heat sink). Pull out the RAM and put it back in. Pull the graphics card (if it’s a separate card and not built into the motherboard) and put it back in. Use compressed air to blow the dust out of everything. Obviously, this should be done by someone who knows how to handle such things without damaging the computer (i.e. you should follow standard anti-static precautions, etc).
Is the power supply fan spinning when the machine is running? Is the CPU fan spinning? Is the case fan spinning?
Have you modified any system parameters like the pagefile size?
Those are the symptoms I get when I have insufficient virtual memory to handle all of the running processes. The machine will slow to a crawl and then eventually freeze rather than crash.
I can go into detail on how to investigate if you need more info.
As long as you have the pagefile set to “system managed” after each crash/freeze it should increase the pagefile size - as long as you have enough space on the pagefile drive - meaning that the crashes should gradually become less frequent.
I keep a RAM defrag tool running on my taskbar that shows current memory usage. While you are maxed out in the amount of memory your OS can use, it may all be in use due to leaks from the programs. Having a RAM monitor going should tell the story. Of course, you can do the same with system tools for test purposes.
I will try running memtest for a longer period of time, perhaps while I am away at work.
As for PSU, I am not sure; I have to open up the computer to check. I will try to get readings of the various temperature.
The freezes are unpredictable. The computer would freeze while entering Win 7, or browsing website, or would not freeze for a couple of days. It’s the same with games, but just more frequent. I could play an entire day without problem, or on a bad day, the computer would just keep hanging up.
There is no sign of system showdown. Everything is responding fluidly when the mouse cursor would just stop moving. In games, frame rate would be smooth before the screen suddenly blacks out and the sound just keep stuttering, repeating itself.
Hence, I am not sure if it got to do with virtual memory, as the problem also occurs on start-up.
I check up my RAM box and it has a lifetime warranty (Xtreem Dark), so I will try getting an exchange, in the process while verifying if the RAM works. If the RAM is clean, or there is no improvement after the exchanging and reseating all the other components, I guess I would take a stab at the PSU and GPU.
Speedfan - will show you temps for many system components - although some will not always be clear - for example temp1 can apply to different things on different machines.
Don’t run the stress tests until you are sure your temps are in range. I have a Q9450 running at stock and normal temp for core0 is 45Celsius with ambient temps of 78Fahrenheit. If substantially over 50C with no load, you need to check the heatsink on your CPU. Dusting out the case with some compressed air is also a good a idea if you’ve never done that before. focus on heatsink and memory chips and PSU
Intel Burn Test - will stress test your CPU and memory. Set on maximum and run at least 5 times to get a quick and very, very stressful test of your system
Prime95 - another stress tester - not as stressful as IBT.
You definitely have a heat problem with the CPU. I’m not sure if you have a PSU problem as well or if you are getting a lot of thermal noise in the measurements from he CPU getting too hot. I’d fix the heat problem first then run the tests again.
PC power supplies use switching regulators, which are a lot more efficient than linear regulators but they do tend to be a bit noisy electrically. I wouldn’t expect to see a perfectly flat line. They should be a lot cleaner than what you have though.
Rule out software related issues before stressing the hardware.
Install the latest graphics/motherboard drivers/DirectX.
See if there is an option in the BIOS that automatically tries to turn off the computer if it reaches a particular temperature, and try turning that option off.
Lastly, try turning off hardware acceleration for your graphic card in Display>Properties, then use the computer for non-games purposes and see if it hangs. If it doesn’t, drivers might be to blame.
On most systems like that there should be at least 5 fans running - One in the front of the case pulling air in, One on the back of the case exhausting air, one on the graphics card, one on the CPU, and one in the power supply (also exhausting air from the case).
Were I trying to fix this the first thing I would do is buy the biggest heatsink/fan I could find for the CPU and use Arctic Silver or similar as thermal compound (use acetone to remove compound already on the Heatsink/fan you buy, if it is already there). Then I would check to make sure that I had the biggest possible fans in the front and back of the case, and that all possible fan mounts were filled. If the front case fan is 80mm or smaller I would probably try to find a way to mount a 120mm fan if the case allows it at all - sometimes you just need to take the front off and drill the proper holes for a bigger fan but it otherwise fits. If all fans are filled and bigger fans aren’t possible you can get fans that move more air, at the cost of more noise - places like http://www.svc.com/ sell those specialty fans.
