Comp Reset Problem

[ViciousXUSMC]

Been around my main forum I browse at work and evga's forum to no avail so last stop the smartest guys around! DFI street crew! See if you can help me out guys.

 

Well here is the thing, I built a new system specs are as follows

 

Long post bear with me it needs to be long to give sufficiant details so nobody just has to ask later.

 

AMD Opteron 146

Geforce 7800 GTX (evga)

DFI Lanparty UT Ultra-D

Aspire 600w Darkside PSU (38a on 12v rail)

512 x 2 OCZ Platnium DDR533 Ram

and misc opitcal drives/hard drive not related to this subject.

 

Now I am a experienced computer user and consider myself "advanced" when it comes to trouble shooting and building things.

 

I had that opteron doing 2.8Ghz at stock voltage no problems prime95 stable overnight, superpi 32m stable, and max temps where reading about mid 40's I had the cpu going into 3.0Ghz+ stable also but heat was too high for stock cooling and I woudlnt want to leave it like that (hitting close to mid 50's)

 

The ram is rated DDR533 but its the good stuff from OCZ I have it 1:1 with my cpu @2.8 overnight memtest86 no errors after I tweaked a few bios options. (drive strenth, data strenth and some added voltage)

 

So we have effectivly proven my OC cpu and ram stable.

 

I played like this for about 2 weeks no problems till one day (I belive when I tried to put in some WDM drivers with the 91.85 forceware) My games started crashing at random eather back to desktop or the computer would reboot, upon windows reboot I got a "device driver error" message caused the crash and when UT2004 crashed to desktop I also got somthing that looked like a driver error message but thru much googling and everything could not find the cause.

 

Tried changing the forceware drivers to 81.94 and 81.95 or what ever it was, both times went thru the totall unistall process using driver cleaner pro. No fix to the problem.

 

So few days ago as a last resort I just reinstalled windows its a new system so not much to lose.

 

Ahah! fixed no more crashes after playing all day, but I am lacking my speed playing at 2Ghz and DDR400 when my system does 2.8Ghz DDR560 easy.

 

Bring my OC back up and 30 min into a game crash, then since then multiple crashes all resulting in windows rebooting, and once or twice just application freeze.

 

This time tho no error reports, just reboots and thats it. So nothing refering to a driver problem.

 

Again tested memtest86, super pi, and prime95, and sandra burnin, temps, voltage, cpu, memory all check out fine, no errors nothing.

 

Only thing to note is temps are up to almost 49c vs the old temps of 43c, thats just due to higher room temps and maybe some dust in the heatsink already. But that temp is not even close to causing error's

 

Video card temps never go above mid 70's and my video card is near 3 fans one intake in front and 2 intakes on side.

 

Using heat monitoring and voltage monitoring everything checks out.

 

Thing is in effort to fix the issue I lowered down my OC step at a time and got to a point at 2.5Ghz where no crashes occur.

 

What I have my friends is an issue one that nobody has been able to answer yet and I cant seem to fix.

 

What can be causing these issues? I am starting to think video card because its the only thing I cant test and im only crashing in games, temps get higher on the cpu in sandra or prime than they do in a game, video card temps are cooler than most due to my 10 fans in my case, voltage readings are fine and my psu is rated for 600w 38amps in 12v rail.

 

and it was stable at first for a week or so, then those driver errors came about, after the windows install now I get crashes when OCing past 2.5Ghz

 

Who has a logical answer and can help me fix it?

 

More info on the subject-I plan on breaking out my voltmeter soon and seeing if voltage drops directly from the meter rather than software,

 

My OC settings are fine like my HTT multi and ram divider ect so dont think that my OC settings are an issue, and the extensive testing proves that they are ok.

 

Most logically it would be a heat issue or psu issue but both check out with testing.

 

Windows I just reinstalled on its partition didnt format the partition first maybe still somthing in the partition causing errors?

 

Thanks for reading it all and thanks for the help!!!

[ViciousXUSMC]

Update:

 

Got off work and hooked up my digital multi meter to my video card/12v 5v lines to see actual readings of PSU instead of the mobo readings.

 

got these results

12v mobo reading 11.96 actual 12.30

5v mobo reading 5.08 actual 5.13

 

in full load (games, bootup, prime95 ect) never saw it drop past 12.20 on the 12v and it stayed super steady at 12.30 for the most part.

 

5v line never droped below 5.08 so that means the PSU is good!

 

I have had no issues for the last 2 days @ 2.5Ghz OC and never have.

 

So just cranked it up to 2.7 and within 1 minute in a game it crashed on me and rebooted windows! the temps had not even had time to go up to upper 30c's yet! and the multi meter was hooked up and votage was good the whole time.

 

So right now I can 99.9999% say its not the PSU or HEAT. and the fact that it was 100% stable just last week before that driver error stuff has to mean somthing.

 

I really need some expert advice. It looks like the mobo must be messed up or somthing, the video card shoudnt have any diffrences if my cpu/ram is oced or not since its voltage remained the same.

 

any ideas anyone?

[JiminyC]

I had that opteron doing 2.8Ghz at stock voltage no problems prime95 stable overnight, superpi 32m stable, and max temps where reading about mid 40's I had the cpu going into 3.0Ghz+ stable also but heat was too high for stock cooling and I woudlnt want to leave it like that (hitting close to mid 50's)

Trying to run that cpu at stock voltages at those speeds is causing your problems. That is an easy one to see. Try increasing your vcore. If it is the "HE" it's rated to 1.5, if it's not I believe its rated to 1.4. You may also want to leave your ltd/ht set to auto as the opterons are only 800Mhz HT. Again, I know many people who have oc'd the 146's and 170's to amazing speeds from 2 GHz but it takes more voltage. Funny it even booted without vcore bump. It's amazing your running 2.5 at stock voltages. In any event, you say you are experienced, then experience should have taught you that oc'n typically requires more voltage. You did that to your ram but not your cpu.

 

GL

[ViciousXUSMC]

Update!! still need help no one reply yet

 

There is a more advanced way to trace problems thru administrative tools in windows xp, when the computer has a error or crashes it leaves data of what happened in my case nothing is there or it just shows an error happened with no explanation.

 

But there is yet another way to find out whats going on (at your own risk!!)

 

In Control Panel >> System >> Advanced >> Startup and Recovery there is an option always on by default that makes your computer automaticly restart when a error occurs.

 

I turned this off to see what exactly is happening when the computer crashes, that blue screen that I see for .0002 seconds when the computer reboots? well now It stays up so I can read it!

 

Windows had an error and had to close down....

 

Error in nv4dispaly.dll so there ya go! thats the problem somthing with the video card. But if its only happening when I oc past 2.5ghz and not 2.8 and it was totally stable just 2 weeks ago that means eather I do need to do a total format/install of windows and see if it fixes it as opposed to the reinstall over windows without format like I did last time.

 

Or the mobo is bad and is not working well with my video card when It's OCed to those limits.

 

Temps, and PSU are all ok and tested thru various methods to confirm this.

 

Now just need sombody familiar with nv4display.dll errors to maybe point me in the right direction.

 

Again temps and psu are not the issue, and I only get crashes in games when Oced higher than 2.5Ghz its totally stable at 2.5 but crashes on me at 2.6 or higher.

 

Please guys shed some light on this for me.

[ViciousXUSMC]

Oops I did have one reply but its incorrect ^^; Jiminy my opty is rated for the full 1000Mhz HTT speed its a socket 939 not 940 like the old ones and if you did the research there are hundreds of opteron users with 2.8+ oc's at stock voltage and running stock speeds with as little as .9v!!!

 

the idea that I have run sandra burn in, prime95, and superpi overnight with no errors alone prooves the OC is stable.

 

I infact had it stable for 2 weeks its when I started to install a WDM driver that I had issues it seems.

 

I am experienced too, I have OCed many computers and aided in hudreds more since I am a regular poster in other forums. My opty with added voltage was stable at 3.2ghz but heat was too high for my cooling.

 

I can 100% assure you the OC is stable on the CPU and have also checked the psu with my multimeter so the power supply is not an issue.

 

Thanks for trying to help but post with incorrect info will only lead to making it harder for me to find the answer to my problem.

 

If you read the post above you can see I traced it down to a nv4display.dll error but dont know why I get that only when OCing. I am starting to think its a mobo problem not working well with my video card when its OCed to that point.

[ViciousXUSMC]

Oh guess I didnt put in there that I was running super steady for 2 weeks at 2.8Ghz 1:1 and started getting errors out of the blue refering to a "windows had to shut down to a unknown driver error"

 

but reinstalled winows and got rid of that error, now I just get General Protection Faults in UT2004 if it crashes, or if windows reboots I get message that error occured in nv4display.dll.

 

So not sure what that entails but do know for sure other than in game crashes system is stable in all test, temps are ok, voltage is ok, and it was stable once before. It sounds like maybe my mobo is dying on me :'(

[ViciousXUSMC]

still no answers? I know its confusing because I couldnt figure it out myself.

 

All the "standard" problems check out to be fine.

 

temp good

psu good

memory good

bios setup good

drivers up to date

 

errors include

 

general protection fault (ut2004 crash)

nv4display.dll error (if window crashes in a game

and today for the first time saw

envy24.dll (maybe .sys cant remember) error when window rebooted in a game.

 

so the errors seem kinda random but only happen when I overclock past 2.5Ghz

 

again it was stable at first just suddely started giving problems.

 

any ideas pleease?

[lowboy]

Why mobo and not gfx card?

Have you tried dropping the overclock on your 7800 to see if you still have the problem.I know your thinking well it only happens when above 2500 on cpu but the two interact. And you still may think about bumping your cpu voltage abit to see if it helps,it doesnt hurt to try instead of having your mind made up it one thing.

[ViciousXUSMC]

Again already ahead of everybody, already tried going from the 1.36v I use and past all test with up .02v at a time all the way to 1.57v and still the problem persist.

 

and yes I already lowered video card speeds to stock to eliminate that out of the possible problems list.

 

I havnt had time to fully test it all out yet but for about 2 hours today I had no problems at 2.7-2.9Ghz in games when it would crash in less than 10 or 15 min usually so looks like I fixed it.

 

But I cant belive nobody else was able to come up with a answer and I ended up finding it myself first.

 

But atleast its fixed.

[Kabong]

Hi,

 

A friend of mine has a problem like yours. But in his case, the system reboots even at default speeds of the CPU (no OC on CPU or GPU, all default). His system (Opty 146 CABNE) does Prime for 12h at 2,9Ghz @ 1.4v, passes 14h of memtest (default or with OC), passes all Sandra tests, but reboots, or there is an error when a 3D application is running. It doesn't matter if the system is overclocked or not, it always reboots with 3D applications. And yes, we have tried another graphics card and another CPU, and another DFI Ultra-D.

 

We have tried everything, an after testing with another memory kit (OCZ VX) the problem is gone. But the other mems he was using (GSkill FC) don't produce errors in memtest, and work OK in a DFI Infinity nF4. No errors or reboots, same CPU, mem and graphics card. :confused:

 

So, could you please let me know how did you solve your problem, so I can try that solution in my friend's computer ?

 

Regards,