Sign in with
Sign up | Sign in
Your question

Experiencing random crashes

Last response: in Systems
Share
June 8, 2011 5:32:35 PM

PC Parts

-Asus p7p55d EVO Mobo
-8GB Ripjaws DDR3-12800 Memory (4 Originally, 4 Added later, crashes occur with 4gb of old memory, or 4 gb of new memory, or with all 8 in so i didnt think this was the issue)
-Saphire ATI 5850 (Now XFX 6850, i opened the 5850 up for cleaning, and broke one of the screwing points taking it apart, i figure i couldnt sent it for warranty because of this although it works, i believed it was what was causing my crashes, but after swapping new card in im still having the same issues.)
-2x 7200RPM 1TB Drives (not raid, 1xSeagate 7200.1, 1xWD Black)
-1x 5400RPM 750GB Backup Drive (WD Green)
-650W Thermaltake PSU
- Core i5-750, Stock Heatsink (Have two, swapped because it could've been the issue but it happens with either 750 i use, so i didnt think it was that.)
-Antec 900 Case

Issues I am Having

Ever since i built the system in January of last year i have had some issues. It used to Blue Screen with Memory Management or ATIkmdag.sys errors, so i purchased new memory thinking i had bad ram and swapped it out, the crashes still occured. I did more research on the atikmdag.sys which was the crash i had the most, and found out that different people said they fixed it in different ways so i tried it all. Some said it was the heatsink did not have proper seating on the processor and it was overheating, so i made sure i applied the thermal paste correctly, when the crashes still occured, i swapped the processor out with a backup i had, and still the crashes occured. Others blamed it on ati drivers for the 5850, so i deleted the catalyst center and used windows stock drivers, and i still had the issues.

Now I no longer get BSODs the computer just shuts off and the only two things i havent replaced are the mobo and the PSU, so i feel it has to be one of those two, but i cant just go out and spend that much money so i need some guidance on what i can do to narrow down exactly what is causing my issues.

Tests I have Done

- I have only used windows memory diagnostics tool to test memory but it found no errors.

- An application named VideoCardStressTester found errors and issues with my 5850 in less than two minutes after running the test and told me the card had issues so i went and purchased the 6850, which passed the test multiple times, but i experience frame rate drops in games and the computer is crashing randomly, but only during activities like photoshop or games.

- I also ran Prime95 for about an hour and the computer did not crash, but my processor peaked at 85 C which i think is a bit high (Measured on CoreTemp), under 5-10% load its usually around 45 C.

- I have tried both 32 and 64 bit fresh installs of Windows 7, the both crash, but i no longer get BSODs the computer just shuts down.

I will probably run a legitimate memory tester when i sleep tonight, to rule that out, but any guidance an advice on how exactly to find what is casuing these random crashes would be appreciated.
a b B Homebuilt system
a b } Memory
June 8, 2011 5:49:51 PM

1. Use Memtest86+
http://www.memtest.org/

2. USe OCCT's GPU test to monitor temperature AND voltage.....voltage anomalies can cause sustems to go haywire. ATX standard is +/- 5% .... moderate Overclock, I look foir 2-3% ..... hio OC's I look for < 1%.

3. Have you checked Event Viewer for error / warning messages at time of shutdown.

4. What are CPU and GPU temps ? Again .....OCCT will monitor, record and graph this for you.
m
0
l
June 8, 2011 5:57:01 PM

It said my GPU does not support CUDA, so i cant run the GPU memtest, should i just run the GPU OCCT? and ill post some results by tomorrow on the tests i run. Thank you for this utulity.

Edit: OCCT could not perform the first test i tried (Power Supply) because 2 minutes in the CPU got too hot.

I have also looked at the event viewer and i have a TON of errors in there dating back to March. But all the recent ones that look like crashes have empty fields, while older ones say the bug that caused the crashes but recently that spot is empty. I can post the log in any needed format.
m
0
l
!