stable for 28 days. not stable enough.

[CoinOperatedBoy]

Originally posted by: VirtualLarry
Passes prime95 24hr test just fine. Haven't run OCCT:linpack, but I think the temps would cause thermal shutdown.

Surprised nobody latched onto this yet. Linpack does generally produce higher temps than prime95, but if you're already near throttling, your chip is running too hot. You say you're 16-20C away from TJMax... when? During Prime95? Not idle I hope. Either way, that's a little high if you ask me.

I think you should attempt your stock speed experiment again. I don't see how that would cause a "funny noise" in your hard drive. In fact, keep a close eye on your drive -- maybe it's failing and that's what's causing your instability. Check the event log for errors if you haven't.

Other than that, random rebooting with no blue screen does sound like a power problem to me. It is odd that you would be experiencing issues with multiple systems in the same location. Your PSU seems fine, so I might suspect the power coming from the wall. Does your UPS include any kind of line conditioning?

Otherwise... 28 days uptime for a home computer isn't that bad. I know this tends to be an enthusiast forum and we may strive for the best stats, but is that really a big concern if the only issue is an infrequent reboot? If hardware or data seems to be affected, that's another story.

 

[VirtualLarry]

Originally posted by: CoinOperatedBoy
Otherwise... 28 days uptime for a home computer isn't that bad. I know this tends to be an enthusiast forum and we may strive for the best stats, but is that really a big concern if the only issue is an infrequent reboot? If hardware or data seems to be affected, that's another story.

Well, the only problem is, I'm running MagicJack on this machine, so if it reboots while I'm away, then my home phone doesn't work until I get back home to fix it.

Edit: It was rebooted when I got back. Things seem to be getting worse.

 

[Thor86]

Memtest stable does NOT equal Windows stable.

You need to learn to use Prime95 stability testing properly:

http://en.wikipedia.org/wiki/Prime95

 

[ashishmishra]

Are these settings disabled? They caused the same issue for me once, they might be labeled different for you since I had an Asus P5Q Pro

DRAM Static Read Control: Disabled
DRAM Read Training Enabled: Disabled

 

[heyheybooboo]

Originally posted by: VirtualLarry

Originally posted by: CoinOperatedBoy
Otherwise... 28 days uptime for a home computer isn't that bad. I know this tends to be an enthusiast forum and we may strive for the best stats, but is that really a big concern if the only issue is an infrequent reboot? If hardware or data seems to be affected, that's another story.

Well, the only problem is, I'm running MagicJack on this machine, so if it reboots while I'm away, then my home phone doesn't work until I get back home to fix it.

Edit: It was rebooted when I got back. Things seem to be getting worse.

I don't really know what the problem is but I'll toss this out there.

I like my MagicJack. Works great for the most part. But I have noticed a slight uptick in aberrant behavior from the rig (Vista on an NF6150/430) after the installation. I'm going to try to launch it in Win7 in hopes of more smooth sailing overall (VMC plays well with the tuner but W7MC just seems much better) ....

 

[Zap]

Originally posted by: VirtualLarry
Well, the only problem is, I'm running MagicJack on this machine, so if it reboots while I'm away, then my home phone doesn't work until I get back home to fix it.

Is that device incapable of starting automatically with Windows?

 

[VirtualLarry]

So I tried doing everthing, +0.2v DDR2, +0.1v vFSB, +0.1v vMCH. Still rebooted after a day or so.

So now I've removed those overvoltages, and am testing with an overclock of 2.8Ghz (8 x 350FSB), rather than 3.2Ghz (8 x 400FSB). I'm curious where the cutoff line in terms of stability is.

I've come to the conclusion that something is simply flaky in my system, probably the mobo, in terms of long-term stability. Too bad. I try to build computers that are 24x7 stable, but this one just isn't making the grade.

 

[VirtualLarry]

Alright, so at 8 x 350 = 2.8Ghz, I've had an uptime of 31.75 days. So I'm going to call that setting "Stable".

Next step is to try the same overclock, but at 400FSB, so 400x7, and see if that reboots.

 

[Idontcare]

I like this long(er) term uptime/stability study. Please keep us posted as you continue to generate data points.

 

[Gillbot]

Run linux on it and see if it stays stable past the 28 day mark. I used to have a few servers that had a similar issue and I switched them over to Red Hat, well after about 4 years of total uptime without a reboot, Linux had me sold. I put windows back on for grins and sure enough, crashed after about 30 days or so.

 

[WaitingForNehalem]

The Linux kernel can run with bad RAM using the BadRAM patch.

 

[imported_Scoop]

I have a solution. Buy a god damn chip that runs @ 2.8Ghz stock

Now if that ain't stable, something is wrong.

 

[Gerbil333]

I'd blame the ThermalTake 430W. You're obviously experiencing power quality (PQ) issues. I don't like to blame instability on PQ problems, but it can happen. And it's definitely the cause of the reboots.

By the way, I own the same power supply. It's in a box in the closet for a reason. Throw in a nice, new Corsair PSU and report back.

 

[VirtualLarry]

I've got an Antec Basiq 500W waiting in the wings for an upgrade, but I want to finish some of my testing first. (Even if it does take 40 days to test.)

 

[VirtualLarry]

Ok, so I waited for 801hours to shut the PC down and test a different overclock setting.
It was stable for 33.x days at 8x350, so I then wanted to try 7x400 (both 2.8Ghz).

What a mess!

I forgot that I had installed AIM 6.8 (?), so it autostarted when I booted the PC. MagicJack also autostarts. I manually start the SeventeenorBust service.

Well, I attempted a video chat over AIM with my friend. It froze up a few minutes in.
I rebooted, and requested another video chat session. Before the video cam had fired up, it froze again. So I unplugged the webcam and my headset, and tried again, just with an AIM conversation. Froze AGAIN.

So at this point I was confused. I rebooted, and this time I didn't start SeventeenorBust. I managed to have a 20min cam session with my friend sucessfully.

I then shut down AIM, and then started SeventeenorBust to see if it would freeze again. SoB was acting wierd. Client #2 would only reach 5% CPU usage (according to CoreTemp).
Clicking on that window would result in nothing happening for quite some time. It was as if one of my CPU cores was being heavily throttled for some reason. If I stopped block processing on Client #1 (CPU load on Core#0 going from 99-100% down to 1%), then Client #2 became occassionally responsive.

I clicked on Client#2's window and dragged it around a bit, and then I got this:
Blue screen: Stop 0x000000E3 (0x89602980,0x88EB98C0,0x88F58C58,0x00000003)

I'm going to try 8x400 again, and see if I get the freezes.

Edit: 18 min in, and it's still stable at 8x400, and no strange throttling.
Edit: Tried 7x350, to see if perhaps the 7x multi is flaky, but in 10 min, no freezes, and SoB isn't throttling.

Edit: I said screw it to attempting to characterize the various failure modes, and replaced the ThermalTake 430W with an Antec Basiq 500W. Now testing at 8x400 again.

Edit: Darn it. It rebooted overnight with the new PSU at 8x400.


So I tried 6x400, seemed to run alright, although when I first entered that setting, the PC did the double-post and reset to default clocks, so I had to re-enter the 6x400 info again. It did that the first time that I set 7x400 too.
It didn't freeze after 7 minutes.

Curiously, when I restarted to enter the BIOS, under my Advanced BIOS options, the EIST option was missing, and in a different place, an Option for "CPU Multithreading" was available. It purported to disable all but one CPU core. I restarted, and the "CPU multithreading" option is now gone, and my EIST is back. Strange stuff. I'm going to try to flash to the newest BIOS, F13d, from my current F9.

Updated to F13d, now I have C1E (was always there before), EIST, AND "CPU Multithreading" options.

Noticed something else, it seems that with the new BIOS, microcode patch level is at "A3", before I think it was at "A1", so I'm crossing my fingers that some CPU errata was fixed, and it was that errata that was causing my machine to reboot. (After all, Triple-fault on a CPU results in a reset, I believe.)

Edit: I decided to perform some long-term stability testing on this PC's twin brother. Also an E2140 @ 3.2, on an identical P35-DS3R 1.0 mobo, with identical RAM and Antec Basiq 500W PSU. I had, however, upgraded the graphics card to an VisionTek 4850, and it has a 640GB WD HD instead of a 320. It also has Windows 7 beta 1 64-bit on it.
When I powered the machine on, it did like 6 power-on/power-offs, before it came up and had reset to default clocks. I re-did the OC settings (identical to the previous PC in discussion), and it booted, double-booted, and reset the settings. So I went in and did +0.2v vDIMM, +0.1v vFSB, +0.1v vMCH, +0.1v vPCIE. It then booted with the OC settings properly. Oh yeah, BIOS is still at F9 on this new machine.

I installed SeventeenorBust (small bug in Windows7, with the service install, no UI shows up), after first installing and configuring as an app. So now I'm just going to let it run, and see how many days it lasts.

 

[Gillbot]

Fresh install of all software and apps is the only way you'll know for sure.

 

[elconejito]

I don't have anything to add other than I am extremely fascinated by your search for stability.

 

[VirtualLarry]

I went searching for this thread, to link to it in another thread, and I read it, and it seems I never posted a conclusion.

I gave up running one of the E2140 @ 3.2 rigs at 8x400, and instead, ran it at 8x350. The other rig continued to run at 8x400 successfully.

I later at some point upgraded both P35-DS3R v1.0 rigs to Q9300 CPUs, and that's the way that they've been running until the present time, at 7.5 x 400. (3.0Ghz)

No issues that I can see. Oh yes, they now run Windows 7 64-bit SP1 HP too.

I basically figured that one of the E2140 CPUs couldn't run at 400FSB reliably, some of them just didn't like really high FSBs. But it clearly wasn't the mobo, because my Q9300s run at that speed all day (month?) long.

 

[Piroko]

My first thought was that your UPSs probably were not producing sine waves and your PSU might trip over it occassionally, but guess I would've been wrong

 

[bononos]

Is it rebooting because something went haywire or is it a (silently) scheduled thing that windows 7/8 doesn't tell people about?

Could it be the UPS that has monthly reboots set up in its firmware/software settings? That was the case with a server in my office than always rebooted every fortnight until we figured out that it wasn't anything to do with the hardware or OS on the server.

 

[BallaTheFeared]

The problem is probably related to the cpu not being able to handle 400 FSB, try lowering it to 350.

I hope that helps

 

[T_Yamamoto]

This issue has long been resolved. He just wanted to post a conclusion. (Always check the original post date y'all)

 

[blackened23]

28 days later and the CPU becomes unstable? It's a prelude to a zombie apocalypse?!