Just last week, I started my 2nd home-built system. Primary purpose is gaming/3d animation rig. After putting it together, and trying to play Fallout 3, Sims 3 or even World in Conflict, the system automatically restarts after somewhere between 15-30mins. No slowdown in FPS, stuttering or anything like that, just playing the game and then a half a second of BSOD, then restart. Unfortunetly, upon restart, windows does not give any error message or seem to think there is a problem, and as i said, the BSOD is too fast for me to even read what it says. Now im asking for help in this forum section because ive been told by a friend that the problem is either most likely the PSU or the memory. So i am asking for help and people's opinion and experiences here.
1. Click on the Start button and then on Control Panel.
2. Choose System under the Control Panel heading in the list of results
3. In the task pane on the left, click the Advanced system settings link.
4. Locate the Startup and Recovery section near the bottom of the window and click on the Settings button.
5. In the Startup and Recovery window, locate and uncheck the check box next to Automatically restart.
6. Click OK in the Startup and Recovery window.
7. Click OK in the System Properties window.
8. You can now close the System window.
9. From now on, when a problem causes a BSOD or another major error that halts the system, Windows 7 will not force a reboot. You'll have to reboot manually when an error appears.
B. Download OCCT and run the bnechmarks one at a time.....if the system fails reboot and look in the OCCT folder...there will be a series of graphs tracking voltage , temperature and you can see if anything looks out or ordinary.
If it's exclusively during heavy loads (gaming or benchmarking), it's almost always power (PSU) or heat related (typically CPU/GPU) ... bad memory will typically show up in general use (non-gaming applications crashing)
If you have poor airflow, move your computer to a cold room, open up your case and have a giant fan blowing air over it. That SHOULD eliminate heat as a potential problem. If that solved it, well... there you go. You got airflow issues to work through.
If that doesn't solve it, seriously consider a new PSU.
OK, so...after listening to jacknaylor's suggestion of running OCCT, heres what we've got now...
1) After turning of "Auto-Restart" in system properties, games now simply freeze onscreen with audio playback still going. No BSOD, hard restart required, windows startup still thinking everythings ok.
2) OCCT Testing - Ran a GPU test first, completed without errors, temp topping out at 57 degrees every other minute.
3) Ran a CPU OCCT test next - FAILED, only lasted a minute. ERROR Detected on Core 3. Temps on 4 cores ran between 55 and 65.
4) Ran a CPU Lineback Test next - FAILED, lasted 1.5 mins. "CPU Too Hot!" Temps on 4 cores ran between 80 and 92 degrees. Thing about this is that it only lasted 1.5 mins until the program itself, FROZE.
Now, heres the thing about overheating - I am using the stock fan/sink that came with the processor. (Dumb i know, but i am just piecing this machine together now) The case i am using is a APEVIA X-Navigator ATXA8NW-AL/500 Silver / Black Aluminum ATX Full Tower, minus the old 500w PSU. Now if u newegg this thing, you see the case itself has 5 fans, 1 in front, 1 on the side, 2 in the back, and 1 on top. Now, i would THINK that that would be enough to keep things resonably cool, at least for the general air inside - but am i wrong with this?
In any case, now seeing these reports, i am assuming that the chip is bad and needs to be sent back, yes? So do i send just that back? Both the MOBO and chip back? PSU back? or a combination of things just to be sure?
How stable was voltage ? If voltage is wavering all over the place, that's why your CPU maybe overheating.
Here's what I'd do next if your are comfy in BIOS
I dunno beans about the i5 but **If** it is like the i7.....
1. Learn how to restore all defaults .... make sure memory timings in BIOS match your RAM specs, if not fix.
2. Put CPU multiplier at lowest value
3. Put mem multiplier at 10 so mem speed is 1333
4. Run memtest4.0 overnight
5. If that works, it's likely not your memory....restore BIOS defaults
If that goes OK, I'd say your CPU heat sink install didn't go very well. Well I am thinking that anyway, but might a swell amke sure about memory. I'f you're gonna reseat it, might as well use an aftermarket one I figure. Megahalms Rev B is for 1156, is a top end performer and is very ez to install.
Well Jack i WAS in the middle of writing you a reply, but then i received my first BSOD EVER during internet ops. Error read "Page_FAULT_IN_NON_PAGED_AREA", for whatever thats worth. Checked the bios immediately afterward, temp at 38c and the 12v was the only thing fluctuating at 12.3xx. Then after startup, my Kaspersky security said that 2 "programs?" (forget what they called it) 2 things IN Kaspersky could not be started. Clicked on the icon and then BSOD again. this time "IRQL NOT LESS OR EQUAL". After another startup, here i am....
As i was TRYING to say in my first message, after looking at the original tests graphs, i think something is up with the 12v. Heres the data:
GPU TEST - 3.3v @ 3.44 - solid through whole test
5v @ 1.3 - Solid through whole test
12v @ .8 <->1.3 - Dancing like CRAZY through the test.
LINEPACK - 3.3v @ 3.44 - Took 2 dips to 3.42
5v @ 1.3 - Solid through test
12v @ .8<->1.3 - Fluctuated a lot, but not nearly as bad as the GPU test.
Ran Another set of tests, same basic results as first. I am going to try running memtest but i think im gonna return the PSU first. Anyone think the chip should be sent back as well? (Error on core 3 really worries me if its not heat related) Thanks to all who respond.