Cause of Random Restarts?

[Elfear]

I'm trying to figure out which component is giving me grief. I'm ocing my system and it passed memtest 4 times so i'm hoping it's not the ram (I know I should run it overnight to really rule out the ram). I can boot into Windows fine and it runs 3DMark01-05 with no problems but if I'm surfing the net or messing around on the desktop than the computer just randonly reboots, no bsod or anything. Here are my bios settings:

290x9
HTT: x3
1.50V Vcore
9/10 divider so ram@262 3-3-3-10 1T
2.80V Vdimm
1.60V NB

Any idea what would cause this?

 

[Duvie]

Do you have automatic restart checked in the system / startup and recovery of winxp??? it may be masking a BSOD error...

Otherwise reboots are usually heat or power...

 

[Elfear]

Automatic Restart was enabled. Thanks for the tip Duvie. I'll see if I get a bsod error message now.

Is there a list somewhere that deciphers what each error message means?

 

[Gamingphreek]

Not really, but if you get one [a BSOD] look in the System Log then to www.EventID.com and type in what it prompts for.

-Kevin

 

[BigBadBiologist]

what is your CPU temp? Also, you could probably raise your Vcore a little bit.

 

[Elfear]

My CPU temp is 30C at idle and ~37C under load using Prime95. Those sound low to me so maybe the board is reporting things wrong.

After unchecking the auto shutdown feature the bsods that I'm getting now are win32k.sys and nv4_disp.dll. The second looks like a problem with my video cards but I'm not sure about the first.

 

[cbehnken]

Back your speed down a bit and see if it still doesn't it.

 

[Elfear]

When I'm not ocing then the computer works fine. I'm just wondering which component/OS property might be causing problems so that I can isolate it. If it ends up that my system just isn't stable than I'll downclock some but if the problem is fixable without having to downclock than I'd rather do that.

 

[Jman13]

To me it simply sounds like your system is pushed too far. It could be either your CPU can't run that fast (likely) or your mobo can't run the FSB that fast. You can check your motherboard by lowering the multiplier on your CPU as low as it will go and running at the same FSB. If you still get crashes, you've reached your mobo's limit. If not, then you've reached your CPU's limit. Either way, you're either going to have to up the voltage to try and get it stable, or downclock to a slower speed.

 

[Appledrop]

probably the ram, run a larger divider and see if that still happens. edit: though.. it is POSSIBLE could be power issue.. 2 6800gts are going to need a looot, however quality 520w should be enough, its possible..

but yes. id do the following : keep HTT at 290, lower cpu multi to make cpu at stock speed, and set a ram divider so that your ram is running slow.

set that up over night, looping pcmark 04 or something if it has rebooted, probably motherboard.

if no reboot, keep htt at 290, and cpu multi lower at stock speed, put ram back up to 262mhz

set that up overnight, if it has rebooted, probably ram.

if still no reboot, it is most likely your cpu, although a maxed out cpu will usually just freeze, as opposed to reboot...

if all of them have rebooted, then i say its your power supply ( take out 1 6800gt and repeat the tests to verify)



another thing: could be a problem with sata drive, as you say it only happens when doing disk-activity...

 

[InfoTiger]

Originally posted by: BigBadBiologist
what is your CPU temp? Also, you could probably raise your Vcore a little bit.

Agree.

 

[Elfear]

I'll try pushing each seperate component and see if I can narrow things down a bit. I did test each component individually like Zebo suggests in his ocing guide. My mobo booted at a 360MHz fsb and my ram passed memtest at 280MHz so again, like Zebo suggested, I put my limit at 5% below what those limits were.

I guess I need to stress the components a bit instead of just seeing if the mobo will boot into Windows or if the ram will pass memtest a few times.

Also, how can I test if it's the PSU? I know it should be able to easily handle the load unless I got a bad one.

 

[Elfear]

Ok, I've tried a few things and I need you guys' opinions. I'm getting pretty frustrated with my system instability. The system passes Memtest 20+ times and Primes for 14+ hours with no glitches at the following settings:

288x9
HTT: x3
1.50V Vcore
9/10 divider so ram@259 3-3-3-10 1T
2.70V Vdimm
1.50V NB

I can boot into Windows (75% of the time) and surf the net but try and open a 3D application and I get a bsod usually involving the win32k.sys file. Seems like the problems are almost the opposite of the original ones I posted.

I tried reducing the multplier to 6 to rule out the CPU and it's still unstable. I then tried a divider to run the ram@~200MHz and still get instability in Windows. I've tried upping and reducing all the voltages and a slew of different bios settings with no effect.

I can only think it's the PCIE locks not working. If it is the PCIE locks, could a new bios revision fix that or do I need to RMA the board? Seems like the board needs to go back as a few people have had good luck ocing their boards with the same bios I have.

 

[Xed]

Are you sure the sata port you're using is locked?

on my k8n neo2 only 3/4 are locked, not sure if they locked them all on the neo4

 

[Elfear]

I'm not sure either. I've tried both 1&2 and 3&4 with the same results. One of the three errors I get with the bsods is with the ntfs.sys file, so maybe the sata ports aren't locking either.

 

[Xed]

I'd recommend you make sure everything is locked, my guess is that is the problem.

Haven't read up on the neo4 so I can't really help you do that

 

[Elfear]

I don't know of any way to make sure they're locked except the way I'm doing it (?). My thinking is that Memtest never accesses the hdds and Prime95 only accesses them when I initially open the file, so they both run fine. Neither program really stresses the gpus either. But when I try and open something that really pushes the gpus/hdds(?) than I get bsods.

I'm kind of at the end of my rope here and I don't know of any more testing that I can do to confirm that the pcie/sata locks aren't working. The only recourse I see is to wait for a newer bios to come out that fixes this (don't know if that's possible, nor do I want to wait forever), or RMA the board and hope I get a better one.

 

[Elfear]

Bump.

 

[XRaider]

Elfear,
I had the same problem. I even RMA'd my Enermax 420w PS back to Newegg thinking that was it. I got the Antec True 480 and the same issues came back - periodically. I have my Raptor in SATA3. The machine would not lock up in any benchmark. I would get lock ups in IE, Windows Explorer or even HL2. The only thing that fixed my issue was I used the 5.10 Nvidia drivers and disabled AGP Fast writes in the BIOS.
Try that bro and let us know.

My specs:
A64 3200+
1gig OCZ Rev2 PC3200
MSI Neo2 Plt
74g Raptor
SB Audigy 2 ZS
BFG 6800 GT OC
480w Antec True
WinXP SP2

 

[Elfear]

Thanks for the info XRaider. I don't have the AGP Fast Writes option in the bios because it's a PCIE board but maybe I'll try using the 5.10 drivers instead of the 6.39 Nforce4 stand alones and see what happens. I'd really rather not RMA the board if I don't have to.

 

[donnelly]

Sounds similar to my problem
I have ultra version of your board.

Can you boot into windows every single time, or does it somtimes hang
then reboot?
I found it to be very unstable past 245 fsb, with reboots happening quickly,
at stock its happens less but still does sometimes.

With my old nforce 3 i used to run at 266fsb with no problems.

i used to have Nforce 3 board and it was rock solid
with same components, both msi.


No errors messages or BSOD, just random reboots

Sometimes it will boot into windows
Sometimes it won't
Sometimes it can game for hours
Sometimes it don't

My volts rails are +- 0.2 volts, under load
memory slots 1/2 and sata 3/4

Tried latest drives and msi included
Tried re-install on a normal ATA hard disk

Seem to happen only after installing Nvidia nforce drivers
was booting fine with default windoze ones.
However this is not conclusive since, Id have to -reinstall windows again to test
this, i might have been lucky the first few reboots.

I want it fixed and don't know where the problem lies.

380 watt True power enough? 2 hard disks and dvd writer.
Used to be enoughe for my nforce 3 board and 6800GT rock solid.
Though PSU was getting warm, so i added a second 80mm fan, now
its ok.
Temps are idle about 30 load about 45 which seems fine to me, thats with
fans at 7 volts. 2 X 120 panaflow case fans and 1 X 80mm ys tech on cpu.

I have testing another 430 watt psu on it and problem presistised.

Should i get a lend of another graphics card to test?

Or is it definitly the mainboard, i though my msi neo 2 was rock solid so i stuck with the brand are they not good?

--------------------------------------------
K8N Neo4 Platinum Edition
3500+ winny 1GB OCZ PC4400
Hitachi 250GB RAID
Antec Sonata Case 380 watt Truepower
3d Connect X850 XT PCI-E
Coolermaster hyperx heastsink

 

[Elfear]

They're excellent boards they just need to work out some of the bugs in their NForce4 models.