New build freezing when doing low intensity work, and I can't seem to find the cause

[daishi5]

I built a new computer for my birthday right after christmas, and it has been freezing on me randomly ever since.

Specs:
I5-2500k
Asus P8Z68-VLX
8GB (2x4GB) Patriot DDR 133
120GB OCZ Agility 3 SSD
Sapphire Radeon HD 6950 2GB
1TB Hitachi HDD
Antec 620W power Supply

When the system freezes, the image stays on the screen, and the system stops responding completely. I can only turn the system off by holding down the power button. All lights on my USB devices go dark.

I have tried memtest, memtest86 overnight, OCCT on the CPU and GPU, and linx. All of the tests were passed with no errors, furthermore the system has never frozen while running any form of stress test, and the highest the CPU temp reached was 35 C.

I have also never had the system freeze while playing Skyrim, X3, or Dwarf fortress. The system has frozen while doing large file copies from my old machine, surfing the web, and sitting idle. I suspected some form of power saving mode because the freezes don't seem to happen when the system is doing the most work. I also believe that the power supply can't be the issue, normally I would test it anyway, but I don't have another power supply that I can replace it with for testing, so I haven't reached that test yet.

I have a coworker with an almost identical system build, but he has had no issues, so I don't think it is a "normal" issue that would be fixed with an update, but I can't seem to identify any piece of bad hardware.

I have tried:

• removing one stick of ram at a time, it still froze with each one.
• switching the dimm slots the ram was in from 1A 1B, to 2A 2B (the manual recommends the second setting, but the system still freezes.
• Re-installing windows to the SSD
• Installing windows on the HDD and removing the SSD
• Replacing the motherboard (I forgot to order the motherboard from newegg, so it was the easiest piece to replace, and originally I suspected some form of problem with power saving)
• overclocking the CPU (I know its strange, but I suspect some form of power saving setting, which I hoped OCing might prevent)
• disabling all power saving settings in windows and the Asus configuration tool
• apply all windows updates
• I currently have the video card removed from the system as a test, but it will take a couple days without crashes to tell if I found a cause.

So far, none of this has helped.

My next steps are to flash the motherboard bios, and then buy a new power supply to test that. I don't know if Newegg will let me return a CPU just to try to test if the CPU is the problem, but it seems like a really crappy thing to do to them.

After those steps, I really don't know what I can do. The freeze produces no dump files, event logs, error messages, or anything else I can look up.

A lot of people report freezing issues with windows 7, but there are so many different forms of "freezing issue" that I find thousands of different fixes.

I don't know what else to do with this system. Has anyone seen something like this, or have any better ideas for how I can try to troubleshoot this?

 

[Bubbaleone]

This discussion (rather lengthy) from TechNet may be worth a read: Windows 7 Freezes when Resumed from Idle

 

[daishi5]

Unfortunately, the technet article discusses freezing when resuming from an idle state. My freezes happen while I am working, and the system is completely unresponsive, while the problem described can be fixed with a reboot through ctrl-alt-del, so I don't think I have the same problem.

I discovered I still crashed without my video card.

My SSD's firmware is up to date.

I also crashed after flashing my bios to the most current version.

I did discover that Microsoft released a hotfix for a deadlock condition causing windows to freeze. (http://support.microsoft.com/kb/2575077) I have applied this hotfix, and I am crossing my fingers.

Edit: Still crashing after hotfix. I also enabled manual crashing to try to capture a crash dump, but the system will not respond to the crash command when it freezes. Apparently this suggests a hardware issue. So, the search continues.

 

[fastamdman]

First up, try loading fail safe options / default options in your bios. We need the system at stock to see whether or not if it's the overclock that is causing it.

Also in your post you said and I quote
"I have tried memtest, memtest86 overnight, OCCT on the CPU and GPU, and linx. All of the tests were passed with no errors, furthermore the system has never frozen while running any form of stress test, and the highest the CPU temp reached was 35 C."

This is 100% bull. There is no way possible that your 2500k on any air cooler is at 35c while at load. Download cpuid hardware monitor and run LinX, watch your temps. The system probably isn't over heating, but it's a good idea to know what your actual temps are at load and not idle.

Your issues can be anything from not enough vcore, not the right ram timings, not the right ram vdimm, a bad power supply or a bad windows install.

When you download cpuid hardware monitor look at your 3.3v, 5v and 12volt rails. If anything seems out of the ordinary then the psu could be the issue. When looking at these make sure there is a load AT LEAST on the cpu, but it would be good to load the cpu and gpu. Not with a video game but with a stress testing program. Run furmark and linx together and that is a REAL load on the system.

Also when stress testing you need to run these tests for a long period of time. 5 passes or 10 minutes is not stable. You need to run LinX for a minimum of 2hrs before you can consider a new build stable, plain and simple.

Try my suggests and then come back

 

[daishi5]

First up, try loading fail safe options / default options in your bios. We need the system at stock to see whether or not if it's the overclock that is causing it.

Also in your post you said and I quote
"I have tried memtest, memtest86 overnight, OCCT on the CPU and GPU, and linx. All of the tests were passed with no errors, furthermore the system has never frozen while running any form of stress test, and the highest the CPU temp reached was 35 C."

This is 100% bull. There is no way possible that your 2500k on any air cooler is at 35c while at load. Download cpuid hardware monitor and run LinX, watch your temps. The system probably isn't over heating, but it's a good idea to know what your actual temps are at load and not idle.

Your issues can be anything from not enough vcore, not the right ram timings, not the right ram vdimm, a bad power supply or a bad windows install.

When you download cpuid hardware monitor look at your 3.3v, 5v and 12volt rails. If anything seems out of the ordinary then the psu could be the issue. When looking at these make sure there is a load AT LEAST on the cpu, but it would be good to load the cpu and gpu. Not with a video game but with a stress testing program. Run furmark and linx together and that is a REAL load on the system.

Also when stress testing you need to run these tests for a long period of time. 5 passes or 10 minutes is not stable. You need to run LinX for a minimum of 2hrs before you can consider a new build stable, plain and simple.

Try my suggests and then come back

Well, I may need to be more clear, the Asus sensor suite reported the CPU as being at 35c while I was running the linx tests, when I first boot the system it is around 26c. If those numbers are not realistic, that might mean the sensor is faulty. Is there an actual temperature sensor in the processor, or is it part of the MB? I don't know what a realistic temperatue range might be, but if the sensor is bad, whichever part has the sensor might be my problem.

While I did try overclocking it, it was just a desperate attempt at disabling the power saving settings. The system was not stable when it was first built at stock settings, and I have restored it to stock settings and it continues to be unstable.

What would be acceptable ranges for my power on the different rails? From memory, the 12v seems to be almost exact, but the 3.3 was somewhere around 3.4. I will collect some more accurate numbers when I am home.

 

[daishi5]

Turns out asus monitor is not updating temps when the system is busy, asus monitor reports 28c while cpuid reports 40c while running linx.

The test is currently running, but I had to restart it because it failed with an error in the first 5 minutes when I ran it the first time. Since the system is running at stock settings, what does that tell me? Can I assume the CPU or MB is bad, or could other hardware still be the issue.

Power on the rails
12v: 12.9 - 12.1
3v: 3.4 - 3.37
5v: 5.0 - 5.0

I honestly dont know if those numbers are too high.

 

[daishi5]

Linx tests keep failing, which before it had no problem. I am really confused. I am worried that something has been damaged, like by the power supply, but the numbers are within specs.

https://docs.google.com/viewer?url=http://www.formfactors.org/developer%5Cspes%5Catx2_2.pdf

I am thinking I will get another power supply because it was my next step in troubleshooting, but I don't know if I should be considering replacing the motherboard or cpu.