System crashing due to # of RAM sticks????

[bntran02]

Update:

I started to get additional weird behaviors.

1. Even when "stable" my computer would sometimes become unresponsive. The screen would be off so I can't say much about what it was last doing. It would never wake back up. This usually takes anywhere from 1 to 3 days to occur.
2. After being in this weird state I would have shut the power. It would have A LOT of trouble turning it back on. It seems to help if I leave it unplugged until all LEDs turn off but that inst guaranteed to get it back up and running. All LED's on the motherboard would turn on but that's it.
3. Something I noticed that has also been happening is that none of the devices connected to my PC were giving any indication of powering up. ie: external HDD, monitors, USB drives, etc... Pretty wide spread. I would know immediately when things are good when I hear my external HDD's spin up and the monitor would turn on.

So ultimately, I just bought a new motherboard. The EVGA WTF X99. Problem could have also been the CPU or PSU but I talked myself back to it being the motherboard being the most likely problem. For now, it seems stable with all 4 ram sticks. Windows just finished installing and I am at the stage of restoring my programs. I've gotten farther than I did in two years with my other motherboard. No overclock of anything yet but that'll come later.

 

[R81Z3N1]

It's always best to find out what the problem is, to me it sounds like a PSU problem but you need to isolate the problem and solve before you add more variables to the mix.

If it was a PSU, or memory problem installing it on the new board might show the exact same signs and then you won't be any better off. I would suspect other items before the board, usually motherboards are pretty good usually they work, or not but always try to find the problem.

I would also suggest before you install your new board to verify if it was memory, or PSU or software problem. You would be surprised what a clean install of an OS can do for stability, and solving weird problems.

If you are using all new items a new PSU, and memory on your new board then of course not much need to trouble shoot, but if not it will be a good learning experience for you. I would also suggest doing that first even before you open the motherboard box, you might find a solution and save some funds.

Granted I get how at times we just have to upgrade, I get that but not fun when your new shinny doesn't boot, and you wonder if you now have to send it all back, or do an RMA. Good luck hope things work out.

 

[BonzaiDuck]

Just popping in, doing a quick scan of posts falling short of a thorough reading. Colonel Mustard is an old hand here, begging the Pulp Fiction joke about the tomato family walking down the street.

It doesn't appear to me that you're overvolting the RAM. Further, not familiar enough with the Haswell-E unpublished specs, but I also doubt you've overvolted the processor enough to matter over this length of time, and the OC of 4.3 is ballpark modest.

However. One thing you don't mention about your rig is the PSU. You might have had the processor for as long as three years. I've just gone through some maintenance and re-assessment of our household systems. One 6-year-old 750W "Gold" X-Series PSU would fail half the time coming out of sleep state. A 650W "Gold" G-series (also Sea Sonic) in another system seemed to be causing Stop Code xxxxxx 09f "Driver_Power_State_Failure" revealed through Blue-Screen-View. That system had been reasonably overclocked in respect to both the 2700K processor and the 2x GTX 970 SLI graphics.

The sleep-state problem involves RAM directly as either a symptom or a cause. Testing the RAM in the second system shows the 2x8GB + 2x2GB = 20GB same make, model and specs, all of the RAM was still tip-top.

This month, I'm buying replacement PSUs. Problems gone. Wasn't either of the motherboards -- p8z68-V Pro and p8z68-V Pro/Gen3 are still solid. I'm surprised either that nobody thought to mention a PSU problem-source, or that I'm ignorant or off base for bringing it up.

As I said more generally, I've ordered two Seasonic X-Series Titanium 650W, received yesterday, one already installed and tested. I have come round to the view that buying a $100 second-tier Seasonic is false economy if the Titanium units cost only ~$40 to $60 more.

But I'm just suggesting: look into it unless someone else here says I'm delusional.

And also, I missed your post updating per your motherboard replacement. I, too, thought my problem was a mobo. But -- it wasn't.

 

[bntran02]

It's always best to find out what the problem is, to me it sounds like a PSU problem but you need to isolate the problem and solve before you add more variables to the mix.

I agree, but I don't have spare parts to swap out. I felt like I was at the point of guessing and buying one part at a time and returning if the same symptoms occurred.

If it was a PSU, or memory problem installing it on the new board might show the exact same signs and then you won't be any better off. I would suspect other items before the board, usually motherboards are pretty good usually they work, or not but always try to find the problem. I would also suggest before you install your new board to verify if it was memory, or PSU or software problem. You would be surprised what a clean install of an OS can do for stability, and solving weird problems.

I did try to find the problem. All memory tests passed with ~72 hours of running memtest86+ with memory in various configurations. CPU was always stable on intel burnin tests. How would you even go about testing the PSU? Voltage readouts appeared correct.

I have had the opposite experience with motherboard reliability. For me, replacing the mobo has been the most common solution. Usually with the exact same model.

If you are using all new items a new PSU, and memory on your new board then of course not much need to trouble shoot, but if not it will be a good learning experience for you. I would also suggest doing that first even before you open the motherboard box, you might find a solution and save some funds.

I am debugging a system that has been up and running for ~2.5 years. Only recently did it just started to crap out over time.

Granted I get how at times we just have to upgrade, I get that but not fun when your new shinny doesn't boot, and you wonder if you now have to send it all back, or do an RMA. Good luck hope things work out.

I think there's a miss-understanding as I am coming from a system that has been up and running that isnt a shinny new toy.

But ultimately, I just got a new motherboard and it all works now. But I cannot say definitively if it was the motherboard or reinstall of windows. But I DID choose a motherboard that specifically listed my ram as

 

[bntran02]

Just popping in, doing a quick scan of posts falling short of a thorough reading. Colonel Mustard is an old hand here, begging the Pulp Fiction joke about the tomato family walking down the street.

It doesn't appear to me that you're overvolting the RAM. Further, not familiar enough with the Haswell-E unpublished specs, but I also doubt you've overvolted the processor enough to matter over this length of time, and the OC of 4.3 is ballpark modest.

However. One thing you don't mention about your rig is the PSU. You might have had the processor for as long as three years. I've just gone through some maintenance and re-assessment of our household systems. One 6-year-old 750W "Gold" X-Series PSU would fail half the time coming out of sleep state. A 650W "Gold" G-series (also Sea Sonic) in another system seemed to be causing Stop Code xxxxxx 09f "Driver_Power_State_Failure" revealed through Blue-Screen-View. That system had been reasonably overclocked in respect to both the 2700K processor and the 2x GTX 970 SLI graphics.

The sleep-state problem involves RAM directly as either a symptom or a cause. Testing the RAM in the second system shows the 2x8GB + 2x2GB = 20GB same make, model and specs, all of the RAM was still tip-top.

This month, I'm buying replacement PSUs. Problems gone. Wasn't either of the motherboards -- p8z68-V Pro and p8z68-V Pro/Gen3 are still solid. I'm surprised either that nobody thought to mention a PSU problem-source, or that I'm ignorant or off base for bringing it up.

As I said more generally, I've ordered two Seasonic X-Series Titanium 650W, received yesterday, one already installed and tested. I have come round to the view that buying a $100 second-tier Seasonic is false economy if the Titanium units cost only ~$40 to $60 more.

But I'm just suggesting: look into it unless someone else here says I'm delusional.

And also, I missed your post updating per your motherboard replacement. I, too, thought my problem was a mobo. But -- it wasn't.

The PSU I have is the Corsair HX750i.

PSU was mentioned a few times and when I look back, it's probably just as likely or more likely given the information. It's definitely next on the list to swap out but time will tell.

Right now my system is stable and has been since I swapped the mobo

 

[Topweasel]

This is going to be anecdotal so ignore if you want but, I wanted to give an experience.

In 2012 I built a 3930K SB-E machine. I originally piecemeal purchased a lot of the components. I got two pairs of Corsair Vengeance memory for the system for the 4 channel memory. After a while I ended up having the same problem about 2 weeks after purchase occasionally my machine would start crashing. So I reinstalled windows and the problem continued. Eventually I figure out it is the memory and I removed a stick, problem went away, so I started trying just pairs and running tests and it would pass. But if all four of these sticks were in, problems would develop within hours. I went and checked their forum to see what support was like and figured out they were going to harass me for not purchasing a more needlessly expensive Quad channel memory kit. Digged deeper and figured out that even their more expensive kits would ship decent sized error rates (but still less than a certain cut off). DDR3 was relatively cheap at the time and I could get 32GB for just over $100. So I bought two Gskill Sniper 2x8GB kits. Figured if it didn't work I would be within the return period. That memory has been working fantastically since. My guess is that one or more sticks are faulty to a degree that four of them together places to much pressure on the memory controller to make up for the inconsistency.