Unstable/Intermittent BSOD/No Post/Reboot on new build

[Guest]

Looking for suggestions on where the issue may be:

Specs:
CPU: Phenom II 965
Mobo: GIGABYTE GA-770TA-UD3
Ram: G.Skill 2x2Gb DDR3 1600 Ripjaw F3-12800CL9D-4GBRL
Vid: Gigabyte Radeon 5770 1Gb DDR5
PSU: OCZ ModXStream Pro OCZ500MXSP 500W ATX12V V2.2 / EPS12V SLI Certified CrossFire Ready

After initially putting everything together, it would POST, but randomly freeze/reboot/bsod followed up by *sometimes no POST for several minutes (until waiting a few minutes). Tried installing Windows 7 x64 off USB, wouldn't get past the file loading without above issues. Would have tried mobo/psu then 1 part in at a time, but the no POST was so rare I wasn't sure it was working or not working.

Then swapped RAM out with another computer, installed W7 without any issues. Ran memory tests on the other computer with the ram, no issues. Put the RAM back in my comp, booted fine.

Ran a few benchmarks with CPU-Z/PC Wizard without any issue. Tried playing 1080p vid on one monitor while playing Starcraft2beta on ultra on another at the same time, no issues.

Issue occurred again randomly later. Froze, pressed reset button, booted fine. Ran benchmark with CPUID stuff again, BSOD/auto reboot occurred. Then wouldn't POST, even after powering off/resetting. Waited 5 minutes, powered on, everything seemed fine again, ran benchmarks -- no issues. Ran windows memory diagnostic overnight, no issues.

Any possible ideas on where the issue is would help.. thanks.

Note: Only mobo settings I've played with is RAM settings (after W7 install). (set to 1600 24-9-9-9-9) because auto was at 1333. Just changed it back to auto. BSOD's are seemingly random and never the same.

 

[shortstuff_mt]

What are your idle/load cpu temps? What PSU do you have? A low-end PSU will cause all sorts of stability issues. I don't trust the Windows memory diagnostics software. I would download Memtest86+ from www.memtest.org and run it overnight. Is the RAM installed in the correct RAM slots according to the motherboard owners manual?

 

[Guest]

OCZ ModXStream Pro OCZ500MXSP 500W ATX12V V2.2 / EPS12V SLI Certified CrossFire Ready

CPU temp never reached above 40/50, upper 30s idling. Pretty sure the RAM is installed in the right slots from the manual (slots 0 and 1).

Also, the RAM model # itself is not listed under Gigabyte's website compatibility list. The 1333 one is, so I assumed the 1600 one would work. I'll give Memtest86+ a shot tonight, but not sure if memory would be the issue. It did work in another comp just fine (though only for a few hours).

Edit: The not posting and waiting, then starting makes it seem like an overheating issue too, but the issue didn't occur under high CPU load (benching to reach 100% usage, a couple hours of video/sc2 gaming)

 

[Guest]

Possibly ram issue, auto detected at 1333 24/9/9/9/9, the memory is 1600 24/9/9/9/9 1.5V, but Memtest86+ and Windows Memory Diagnostic running one on each stick (two comp) at the moment, no errors yet.

When running Prime95’s Blend Test, it has caused BSOD/reboot sometimes within seconds. Other times ran without issue for 30min-1hr.

CPU temp is consistent between 54-56 C at 100% all 4 cores during Prime95.

No Post (black screen, all fans on/spinning) is seemingly random.

Anymore suggestions?

 

[capt_taco]

I would definitely set the RAM timings and voltage yourself. That motherboard lists supported speeds as 1866(OC)/1333/1066 ... which I take to mean, if you try to run the RAM at anything over 1333 MHz, you have to treat it like an "overclock," which means entering the settings manually. Even if your RAM's stock speed is 1600 and you're not trying to change anything.

If you just drop it in and let it autodetect, the board SHOULD just downgrade the RAM to 1333 and run it fine, but who knows, it may have a problem doing that. RAM autodetect features are pretty hit-and-miss, and often do weird things for no reason that I can understand. But the fact that you have a default speed that the auto settings may not like seems like the most likely first place to look.

Random freezing certainly would be consistent with that type of problem.

 

[Guest]

Will try that, but been running Memtest86+/Windows Mem diagnostic (this one overnight) without seeing an error at 1333. Sometimes I'll hit the power button to boot and it won't even post (black screen, all fans spinning). Edited previous post with running prime95, but basically is there some reason why something like prime95 running for extended period (hour+) would be fine, and then just bsod seconds into another? I believe prime95 mostly stress tests CPU and some ram?