The most expensive PC I have ever built is the most unstable one and I don't know why

pzduniak

Commendable
Dec 31, 2016
5
0
1,510
Hi,

I really have never had any issues with the PCs that I built. Everything was fine and dandy, stability and temperatures were not a problem. That is until I completed my current build:

MB: ASUS Rampage V Extreme
CPU: i7-5960x with the dedicated X99 EK waterblock
RAM: Corsair Dominator 8x8GB 3200MHz
GPU: GTX 1080 EK
PSU: Corsair AX1500i
SSD: Corsair MX300 1TB

At this point the PC is not even overclocked. Everything packed nicely in the Enthoo Primo with ~1m of thick EKWB radiators.

Here's the history of what I did with the PC:

1) Installed Windows 10 with a GTX 980 and a 512GB MX100 - everything worked fine-ish until the Windows installation decided it's time to engulf itself in flames - explorer stopped working and I had to Win+R everything. I had no issues with stability.

2) Bought a GTX1080, GT210, the new SSD and put everything in the PC. Installed Ubuntu Server and did a double PCI-E passthrough in Windows 10 VMs. Gave up when I noticed that Elgato Game Capture HD60 Pro that I used for production of a live event didn't work with passthrough. Everything was stable-ish, ie. stuff did crash (VMs sometimes froze etc), but given the amount of hacks in place it was understandable.

3) Removed GTX980 and the 210, put the first one in a second PC that I built for LANs, installed Windows 10. Everything worked for a week, after that the PC started randomly shutting down. Updated the BIOS, checked all temps - the max I can get under stress on both 5960x and 1080 is 42C. PC still randomly shuts down for no reason, just reboots.

There is nothing in the system logs, no errors displayed in the BIOS. Just resets during which I lose all my currently opened work.

I am lost. I cannot think of any sources of informations that I could use for debugging the issue. This is my workstation and I want it to be as stable as 4x less expensive shitty mATX PC. Any ideas?
 
Solution
Very strange to have random reboots if its not PSU or temps plus no logfiles. May want to try less RAM modules, 8 DIMMs is a good bit of load on the IMC, even at stock 2133. Just to try, I'd go back to 4, or increase stock 1.2v to 1.25 or 1.3. Even a bit of SA voltage.

pzduniak

Commendable
Dec 31, 2016
5
0
1,510


Took the CS750M from the second PC (6700K, GTX980), let's see. The fact that I did not have remove a 48cm radiator to remove the 8pin CPU cable was quite pleasant :)
 
Watching this as well. PSU test with CS750M is a good idea. Also, are you using XMP? 100 or 125 Strap? I would try running memory at stock 2133 for awhile. My 5820k is very picky on memory straps and DDR4 speed, even when I was running stock. Also, ASUS stock UEFI usually OCs all cores to full turbo instead of one at stock settings. Your uncore at stock 3000mhz as well?
 

pzduniak

Commendable
Dec 31, 2016
5
0
1,510


Because I'm using X99, I'd have to overclock stuff to get more than the default 2133MHz and per the OP nothing is overclocked. So yes, it crashed with RAM being stock. Turbo is set to (whateveristheturboboostof5960x) on all cores, but it has never been an issue.

After around 5 hours of testing it on CS750M it seems stable, so it probably was a PSU-related issue. The shutdowns were quite random and I will be testing it for 2 more days. The PC crashed every day, but not that frequently (once per day, maybe twice). We'll see how it goes.
 
Ok. It does sound more and more PSU related, even with the AX1500i. I suppose even Tier 1 can have defects time to time. Been very happy with my AX1200i for both my X79 and current X99 build. Like you already mentioned, a few days of further testing is best course of action.
 
Very strange to have random reboots if its not PSU or temps plus no logfiles. May want to try less RAM modules, 8 DIMMs is a good bit of load on the IMC, even at stock 2133. Just to try, I'd go back to 4, or increase stock 1.2v to 1.25 or 1.3. Even a bit of SA voltage.
 
Solution