A clock interrupt was not received by a secondary processor

[dvsone1440]

New to building computers... Managed to get it together and works really well.. right up until it doesn't.

I am an architect and built a computer for modeling and rendering. The modeling program I use is a 32 bit program, so I wanted to overclock as clock speed makes a huge difference in the modeling program usability. The computer works great, but.. Its been shutting down on me every few days. Typically when rendering (Maxwell render) and it is using all the cores 100%.

I currently have it overclocked only with the tpu switch.. 4.1ghz

The components are asus p9x79/ intel i7 3930/ corsair dominator 32gb/ corsair 850w 80+/ corsair h100i radiator/ nvidia quadro 4000 / samsung pro 500gb.
Temperatures are very stable.. 73 c was the highest I saw in the 8 hr stress test. I have the latest asus bios installed. Windows 7 64 bit.
I installed aida 64 and get no error messages after running it for 8 hrs.

The one strange thing I noticed was in the power tab it mostly a straightish line graph but every once and a while drops off precipitously. The Power minimums all read -10000.

Again I really don't know what I am doing so hoping to get some ideas of where to start. Googling around there does not seem to be any solutions to this crash that I can find.

 

[dvsone1440]

here is the aida 64 stress test and the administrative event log

 

[Automata]

Hello dvsone1440 and to the forums!

Stability can be a fickle at times, even on a system that you previously thought was "rock solid stable". If it happens every few days ("rarely"), that means you have it nearly nailed down, which is good, especially for your first time. One thing to watch for is when it locks up and what it was doing at that time. For example, if it was doing something processor intensive (rendering, stability test, etc), then you have good reason to believe it is CPU/memory related. If it happens only when the CPU and video card are under stress, but each one passes tests individually, that may indicate an issue with power (such as an underpowered or faulty power supply). Some of the more tricky ones are when the system is sitting idle, as you can't really run a stability test for that, short of letting it just sit there.

The first thing to always check with a new system is that it is stable at stock settings. Have you done this? That way, you know your hardware is functional on the level that is promised by the manufacturer and you can start working on overclocking.

Since you are doing rendering, I assume you will want the same stability that I need for compiling code: absolutely rock solid stable. Tracing a bug back to a stability issue is not only time consuming, but insanity inducing. I suggest not trying it.

I always test my overclocks with LinX, which is a UI to Intel's own code they use for testing stability and benchmarking, Linpack. When running that test, I stop all other programs from running, set it to use maximum memory (push the button labeled "All"), set the number of runs to something absurd, and let it go. I'll let this run around 48 hours, which I would say most people on here agree is absolute overkill for stability, but I'm not taking chances. The longer you run the test, and the more environmental factors you put it through during this test (evening being warmer, for example), the higher your chances of finding instability here instead of during a render (or compiling code!). Be warned though, this is a stressful program and you will see temperatures higher than most other programs. If you see it go above 80-85c, I'd kill the test and work on cooling.

With the message you are getting, I'm guessing it is related to CPU or memory voltage. However, I'd give LinX a run to see if it finds anything. If the system is unstable, it should find it quickly.

 

[dvsone1440]

Been running Linx for a while now with it set to "all" CPU temp has creeped up to 56. I will keep watching it and let you know.

 

[dvsone1440]

Been running Linx for a while now with it set to "all" CPU temp has creeped up to 56. I will keep watching it and let you know.

Steadied out @ 55 degrees running Linx. All cores and memory @100%

 

[dvsone1440]

LinX has gone through 16 tests 9 hrs so far and temps are at 57 degrees. Does this software give you any kind of output that can help analyze? Or is it purely to see if it crashes / overheats?

 

[Automata]

It will be obvious if there are errors. The system will lock up, restart, crash, or the program will detect errors.

 

[dvsone1440]

Ok thanks Thideras,
I am going to let it run overnight and see how it goes. It is on run 22 of 100.
So far no issues with CPU @ 59c and Gpu @ 57c.

 

[dvsone1440]

I needed my computer last night so I had to stop LinX No errors or crashes in 15+ hrs

I ran Prime 95 overnight and just did the furmark for 20 minutes concurrently. No crashes or errors except I did get an asus warning that +12 volts was at 15something. Went away to fast to read fully.
CPU was below 70c and GPU was around 81c pretty steady through the 20 minute furmark test.

 

[Automata]

It is safe to ignore the voltage warning. If it truly hit 15v, you'd be explaining the smell of burning electronics to me, not the results of a stability test. Software voltage monitoring is notoriously unreliable.

The temperatures seem fine, but I find it interesting that they jumped that much for simply adding the graphics card to the mix. Your case may not have enough airflow to keep up when both are going. That isn't likely causing the original issue, though.

 

[dvsone1440]

Ok thanks. Well I needed to run a rendering... so I am back to that. I guess I will just let it go and see what happens.

I actually had forgotten I had the side of the case off when I ran LinX and put it back on when I was running the Prime 95 / Furmark so that probably helps explains the increase a little as well. Should I be concerned about adding more airflow with those temps? I took out one of the drive bays as I did not need it to help airflow. For that matter I could probably take out the other one as well... all I have is a tiny 500 gb ssd held in by a 5 slot drive bay. Must be a simpler way to mount that without all that metal.

I often have maxwell running while I am using sketchup at the same time. Surprised me but the machine seems to handle it fine. Maxwell has the cores pegged but their literature says it only takes what is available and that seems to be true as sketchup runs fine. However when I am working that way maxwell has the cores pegged and sketchup taxes the video card pretty hard, so I will have situations where both are working hard and heating up.

 

[Automata]

If you are loading both the processor and video card up when the system crashes, that may be the cause of your errors and gives us a bit of direction to finding the source. I'd start by attempting to reproduce the problem with the side of the case on. If it crashes, reboot, take the side off, and load the system up again. If it doesn't crash with the side off, then you know the problem is temperature related.