I am at my wit's end...I am going to smash my machine, I swear!

A little over a year ago I bought a dream, that dream came in the form of a 680i based board from Abit...The IN9 32x Max to be precise. Fast forward 1 RMA and 9 months later, and I still cannot get the machine to run stable or reliable. Now, what am I to do? I had spent well over $5000 on the hardware, only to end up with shattered hope and a broken wallet.

Then comes along 780i, and once again my dream of a working machine gets rekindled. I jump on the bandwagon, fork out yet another $250 on a brand new 780i, and guess what? I am left with a $250 more expensive pile of junk, compared with the pile of junk I have suffered with and hated and resented for well overa year now.

Whats wrong you ask? Well, the machine was crashing, hardlocking, and generally acting like a piece of trash since I first assembled it. Now, the exact problem is this: The machine runs 3d applications like crap, and by crap, I mean not at all. It will fire up, run windows (though once in a while it will even hardlock in windows), but the 3d apps are a mystery.

Once in a while I will fire one up and play for 9 hours without a problem. Once in a while it refuses to run anything. It hardlocks, the screen goes blank, then it comes back up with distorted and frozen graphics image. ctrl+alt+del doesnt work: the only recourse is a hardboot (something I don't like to do to my machines).

Now, since this is THE EXACT SAME THING HAPPENNING ON TWO DIFFERENT BOARDS FROM TWO DIFFERENT COMPANIES, what could it POSSIBLY be? Whats the problem? I have tried EVERYTHING! I have tried every one of the 4 DIMMs of ram I have, I have tried 3 different graphics cards, I have tried running different sets of drivers, I have tried tweaking the fan speeds to max to rule out heat crashes, I have tried running the BIOS settings in everything from all manual settings to full automatic everyhitng, and I still can't get my (now) $5500 box to work.

Can some one, for the love of everything that is holy and decent, HELP ME before I smash the living hell out of the hardware and swear off games and computers for the rest of my life?

Tried a differnet psu?
Just took a quick look at customer reviews at newegg for the psu and found this.

Manufacturer Response:We apologize that you are currently experiencing a problem with your 850W PSU. We have found that the 850W PSU has some compatible issues with certain Motherboards. We found the problem that was caused by the min. load limit and voltage regulation, this is not a comprehensive compatible problem with all kinds of system.
This issue has been fixed and improved for the next shipment.
1) We revised the "L6" from 10 turns to 20 turns from the original 850W.

2) How do you know if the version you have is the updated modified version?
Check the box, and if it has the “80PLUS” sticker on it, it means the PSU is the corrected version.

Might not be the cure all for your problems but a flakey psu can cause all sorts of strange issues.

I have never used sli or a nv chipset motherboard so no idea about what is good/bad on those.

The rest of the componets in your sig look solid to me.

Power supply? Have you tried any hardware diagnostic programs such as Micro 2000
hehehe we were writing at the same time Plaicd

Quote:

---

Originally Posted by Plaicd

Tried a differnet psu?
Just took a quick look at customer reviews at newegg for the psu and found this.

Manufacturer Response:We apologize that you are currently experiencing a problem with your 850W PSU. We have found that the 850W PSU has some compatible issues with certain Motherboards. We found the problem that was caused by the min. load limit and voltage regulation, this is not a comprehensive compatible problem with all kinds of system.
This issue has been fixed and improved for the next shipment.
1) We revised the "L6" from 10 turns to 20 turns from the original 850W.

2) How do you know if the version you have is the updated modified version?
Check the box, and if it has the “80PLUS” sticker on it, it means the PSU is the corrected version.

Might not be the cure all for your problems but a flakry psu can cause all sorts of strange issues.

---

Good call, but I don't think thats the case. My friend recently purchased a Coolermaster 1000w Real Power Pro for a new box he is putting together. I borrowed that PSU a day or two ago, right before I pulled the trigger on the new 780i, and it didn't help at all.

As for trying it on the new MoBo, based on what I have seen, I don't think that would make a difference at all.

--Skrips

Quote:

---

Originally Posted by RADCOM

Power supply? Hava e you tried any hardware diagnostic programs such as Micro 2000
hehehe we were writing at the same time Plaicd

---

Nope. I am doing everything the old fashioned way so far. On that note, whats Micro 2000?

Have you run MemTest 86+ or otherwise confirmed your RAM isn't at fault here? Of course, the IN9-32X MAX was a piece of junk from the get-go; its digital PWM was terribly unstable, ran incredibly hot, etc. I think they may have released a second rev IN9 with an updated PWM, but I'm not sure.

As for the 780i, I haven't heard of too many complaints. Assuming you're running the latest BIOS, I'd narrow my testing to the RAM and the PSU.

Quote:

---

Originally Posted by Omastar

Have you run MemTest 86+ or otherwise confirmed your RAM isn't at fault here? Of course, the IN9-32X MAX was a piece of junk from the get-go; its digital PWM was terribly unstable, ran incredibly hot, etc. I think they may have released a second rev IN9 with an updated PWM, but I'm not sure.

As for the 780i, I haven't heard of too many complaints. Assuming you're running the latest BIOS, I'd narrow my testing to the RAM and the PSU.

---

Well, like I said, my friend's PSU didnt help me at all while running the 680i board a couple days ago, and seeing as I am still experiencing the exact same types of crashing, I am going to guess swapping my friend's PSU back into my box is ALSO not going to help now. As for the ram, I tested all 4 DIMMs several months ago when I really started doing some digging.

All 4 checked out, though keep in mind I did not use any diagnostic software. I just ran one at a time and verified that they worked under real world trial.

Although, mayb a test would be in order....but hey, why should I need to test if I am only running them in pairs anyway? More than 1 of those DIMMs couldnt have failed, could they?

well i will say your first initial issue was that putrid abit board.
the 680i was not one of abits better devices. not that i ever saw, but i didnt hang around to watch it get fixed or not.


alot of things can make a pc go bad man, even your mouse,, believe it or not, i have seen mice and keybaords do all kinds of wicked wierd stuff.

next is your overclock. stop overclocking and set it all up on defaults.
bench it, cpu, ram everthing make sure its stable if not adjust your voltages but not too much as defualt settings on that stuff should be pretty close to auto.


what i am going to ask for is what are your current bios settings and what is your current overclock.

you can also lookinto a bad power strip or surge suppressor, if your using a battery backup then dont.
a bad battery backup can cause all kinds of crap to.

start out by IN DETAIL listing the bios setings you are using. and i dont mean just overclocking i mean everything.

take the time and take some pictures and upload them to image shack. or host them somewhere.

let us help you if you are going to ask us for help.

just get an intel chipset thats the easiest fix, having a broken platform for a MB isnt worth sli

Quote:

---

Originally Posted by Skrips

All 4 checked out, though keep in mind I did not use any diagnostic software. I just ran one at a time and verified that they worked under real world trial.

Although, mayb a test would be in order....but hey, why should I need to test if I am only running them in pairs anyway? More than 1 of those DIMMs couldnt have failed, could they?

---

ok, your probably going to be mad at me for saying this but that statement right there pretty much proves you really dont have any idea what your doing.

sorry but thats a fact.

run memtest in windows with the free windows gui tool, or via dos off a floppy or via the bootable cd iso. (and DO NOT download memtest 2.00 it is bad and broken. download memtest 1.70 or something)
then in windows, download OCCT, it is very picky and has a tendancy to find errors much quicker and where orthos wont.

i never liked it until recently and now i use it all the time. you are also guarenteed to have 100% cpu usage on a dual or quad core where as orthos wont, you have to run several instances of it.

run prime if you wish to run 2 or 4 instances of it. they all are prime based programs so no one is really better than the other although some have been reported to be buggy.

let it run for atleast an hour. its default is 30min. which is ok for basic initial testing but since yoru system is so erratic then you will want to run it for no less than an hour or two.

and dont be doing anything while the tests are running.

also what kind of garbage sh|t software do you have installed on your system.

if your system was so unstable i can almost bet you that the windows iinstaller would have blue screened on you a time or two.
if not then the system should be fairly stable, although that is not a guarentee. windows can install fine but the hardware can still be borked.

also windows can install fine but in reality its really messed up so that when you go to run windows all kinds of wacko sh|t starts happening and you blame everything but windows.

it could be 1 million things right now, but not running tests on possibly bad hardware is not wise.

how can yo usay it runs fine for real world things when you are here telling us that your pc is totally screwed up ???

also bad IDE cable, bad anything can make a system go wacko...

Quote:

---

Originally Posted by Lestat

ok, your probably going to be mad at me for saying this but that statement right there pretty much proves you really dont have any idea what your doing.

sorry but thats a fact.

run memtest in windows with the free windows gui tool, or via dos off a floppy or via the bootable cd iso. (and DO NOT download memtest 2.00 it is bad and broken. download memtest 1.70 or something)
then in windows, download OCCT, it is very picky and has a tendancy to find errors much quicker and where orthos wont.

i never liked it until recently and now i use it all the time. you are also guarenteed to have 100% cpu usage on a dual or quad core where as orthos wont, you have to run several instances of it.

run prime if you wish to run 2 or 4 instances of it. they all are prime based programs so no one is really better than the other although some have been reported to be buggy.

let it run for atleast an hour. its default is 30min. which is ok for basic initial testing but since yoru system is so erratic then you will want to run it for no less than an hour or two.

and dont be doing anything while the tests are running.

also what kind of garbage sh|t software do you have installed on your system.

if your system was so unstable i can almost bet you that the windows iinstaller would have blue screened on you a time or two.
if not then the system should be fairly stable, although that is not a guarentee. windows can install fine but the hardware can still be borked.

also windows can install fine but in reality its really messed up so that when you go to run windows all kinds of wacko sh|t starts happening and you blame everything but windows.

it could be 1 million things right now, but not running tests on possibly bad hardware is not wise.

how can yo usay it runs fine for real world things when you are here telling us that your pc is totally screwed up ???

also bad IDE cable, bad anything can make a system go wacko...

---

I agree, with everything here. I have been rather idiotic as to how I have gone about doing this, but there are two things I would like to say about this (thought they are NOT excuses, just points)
1) The parts being as new as they are, I just cant believe they are bad! Not to mention I have NEVER had a part flat our fail on me, and certainly not more than one part at the same time.
2) Yes, I am new to high end hardwar. Really new. I have been building computers since I was really young, but I have never dealt with truely high end parts until now. I am learning as I go, so please cut me a little slack :)

As for the test, surely I will go and start running them now. I want to solve this problem once and for all, and whatever you guys want me to do, I'll try and hoefully we can figure it out together.

First thing first, while I get ready to run prime, could you let me in on just HOW do I go about using memtest if I dont have a floppy drive installed right now? Also, What is OCCT and where can I download it?

-skrips

OCCT is a hardware stressing program, akin to Prime95: http://www.majorgeeks.com/OCCT_d5612.html

Download it and run the 30 minute stress test.

Another question: How the hell do I take pics of my BIOS? :confused:

Quote:

---

Originally Posted by Skrips

Another question: How the hell do I take pics of my BIOS? :confused:

---

A camera.

Btw when was the last time you did a fresh install of the os?

Quote:

---

Originally Posted by Plaicd

A camera.

Btw when was the last time you did a fresh install of the os?

---

Literally I finished installing XP Sp2 early this morning, left the comp on overnight to patch WoW for my "real world testing", then went to sleep. I warpped everything up this afternoon. So, as far as the OS is concerned, its a non-updated vanilla installation of XP with all the latest drivers from all the vendor websites.


PS: I am going to sign off for a little bit to start running those tests. I will be back on as soon as I am done or as soon as things start to really fail on me. wish me luck, and THANKS a bunch for the help! Please keep checking back as I am sure I will be equally helpless as to what to do next.

--skrips

in terms of SP2 a baseline install will run ok but dont forget the nvidia drivers.
i know you wouldnt forget those because you wouldnt even be posting here if you didnt hve those installed ;)

like i mentioned memtest can be run in windows with the free windows based tool or via dos with a bootable iso that you can download and burn.

google is your friend but here it is for you.

windows based memtest http://hcidesign.com/memtest/download.html for 2gb of ram you have to run 2 of them at the same time.
click the button to start the test, and read the error, it will tell you that you can only run like 830mb of ram, or something close to that. click ok, then type in the box 830 then start the test.
once that starts in the 2nd window just click run test and it will use all the ram thats left over.
i ran it to try something a bit ago and it was using 1.86gb of ram.!! its not 2gb but it can only use what windows is NOT using.
thats why the dos version is better, it can use all of 2gb with the exception of only a few meg.

downloadable iso to burn to cd and boot to.
just make sure your cdrom is set to boot before the hard drive, which by defualt it is not.
http://www.memtest.org/download/1.70...+-1.70.iso.zip

do not download memtest 2.0 its totally fuxxored and the guy who wrote it needs to remove it before it hurts someones pc.
it will lock up your pc, it will stall for several seconds upon booting the program. it will in general barely run.
1.70, which is the link i gave you, runs like a champ.

Well, a little update:

I ran OCCT for a total of 30 minutes just to get a test, under the CPU + Ram mixed test. granted, its only a 30 minute run, but still, no crashes. Keep in mind, too, that I am just trying to run the machine at STOCK settings, I haven't overclocked anything.

As for prime, I want to run it, but I thought you neeed to run mutiple instances of it, no? I seem to be having trouble figuring out how to get more than one instance of the program open...

A little help, please? What do I do with Prime95?

Quote:

---

Originally Posted by Skrips

Well, a little update:

I ran OCCT for a total of 30 minutes just to get a test, under the CPU + Ram mixed test. granted, its only a 30 minute run, but still, no crashes. Keep in mind, too, that I am just trying to run the machine at STOCK settings, I haven't overclocked anything.

As for prime, I want to run it, but I thought you neeed to run mutiple instances of it, no? I seem to be having trouble figuring out how to get more than one instance of the program open...

A little help, please? What do I do with Prime95?

---

When you run Prime95, enter "2" in the "Number of torture test threads to run" under Options > Torture Test. Make sure either "Blend" or "Small FFTs - maximum FPU stress, data fits in L2 cache, RAM not tested much" is selected in this window, as well.

Basically, "maximum FPU stress" means the program stresses the floating point performance of the CPU, where a crash or Prime95 giving you a "thread halted" message would signal an unstable overclock or, very rarely, a bad CPU. "Blend" tests everything, as well as a good deal of RAM.

In a non-stability test sense, Prime95 tests integers in the 2^p-1 form to find new Mersenne prime numbers, if I'm not mistaken. Obviously, these complex calculations require stability.

Quote:

---

Originally Posted by Omastar

When you run Prime95, enter "2" in the "Number of torture test threads to run" under Options > Torture Test. Make sure either "Blend" or "Small FFTs - maximum FPU stress, data fits in L2 cache, RAM not tested much" is selected in this window, as well.

Basically, "maximum FPU stress" means the program stresses the floating point performance of the CPU, where a crash or Prime95 giving you a "thread halted" message would signal an unstable overclock or, very rarely, a bad CPU. "Blend" tests everything, as well as a good deal of RAM.

In a non-stability test sense, Prime95 tests integers in the 2^p-1 form to find new Mersenne prime numbers, if I'm not mistaken. Obviously, these complex calculations require stability.

---

OK, maybe im stupid....but under Options > Stress Test there is no option for "Number of torture test threads to run"....maybe I have a different version?

I went here to get the latest version available. Is there another prime95 I don't know about?

--skrips

Nah, that's not the most recent version. I think I have a modified version--just got it off someone here at XS. Either way, if you don't have that option, just run two instances of the program simultaneously, using the same settings for each instance. I'll see if I can attach my version as a zip file for you.

My version is 25.4, BTW.

Better yet, here's the most recent version for you: http://www.majorgeeks.com/Prime95_d4363.html

Google is your friend. :)

Quote:

---

Originally Posted by Omastar

Nah, that's not the most recent version. I think I have a modified version--just got it off someone here at XS. Either way, if you don't have that option, just run two instances of the program simultaneously, using the same settings for each instance. I'll see if I can attach my version as a zip file for you.

---

I thought the same thing, but it won't let me open two instances at once for some reason. Everytime I try, it just brings me back to the instance that is already open.

--skrips

Quote:

---

Originally Posted by Skrips

I thought the same thing, but it won't let me open two instances at once for some reason. Everytime I try, it just brings me back to the instance that is already open.

--skrips

---

See my edits from my last post. There's a link there for you.

Quote:

---

Originally Posted by Omastar

See my edits from my last post. There's a link there for you.

---

Thanks, but I am going to try something first. I figured out how to get two instances of Prime to run at the same time, and I am going to test it that way first with each windows set to a different core.

Then we will try the other tester.

--skrips

*****Update*****

I have completed the following tests:

- Prime95 for 1.5 hours (two instances running, one for each core)
- 30 min quick test in OCCT
- 30 minutes of Memtest


No issues or errors reported from any of those programs, which is almost not good. I am inclined to think maybe its one of my video cards experiencing an intermittent issue? I just cont think of anything else it could be. But at least I am getting an idea of what its NOT:

- CPU
- RAM
- MoBo

Now, are there any stress tester for video that would be good for this sort of thing? I just have this itchy suspicion that its one of my GTX cards...Also, I will snag a camera tomorrow and report on all of my BIOS settings, and maybe we can try and figure out if anything is wrong in that department as well.

Also, I need a good monitoring program for voltages, fan speeds, temps, etc...any suggestions?

Thanks all for the help, I will report back tomorrow!

--Skrips

Quote:

---

Originally Posted by Skrips

Now, are there any stress tester for video that would be good for this sort of thing? I just have this itchy suspicion that its one of my GTX cards...Also, I will snag a camera tomorrow and report on all of my BIOS settings, and maybe we can try and figure out if anything is wrong in that department as well.

Also, I need a good monitoring program for voltages, fan speeds, temps, etc...any suggestions?

Thanks all for the help, I will report back tomorrow!

--Skrips

---

Give ATI Tool v 0.26 a try. It runs an OpenGL (I think) graphics benchmark continually, and if you notice any graphical abnormalities while it's running (yellow or white spots, other discolorations/anomalies) with your cards at stock, you likely have a dying card. You could also give Rthdribl a try, I suppose. Dunno if that's really suited for stability testing.

Why not just pull one card and test each one separately by playing a game or three? It's also possible you have a bad SLI bridge, I guess.

Quote:

---

Originally Posted by Skrips

Also, I need a good monitoring program for voltages, fan speeds, temps, etc...any suggestions?

---

Get yourself the latest beta of Everest, there is a thread in the News section. This program is very good.

Also, I think you should run memtest a little longer, as 30 minutes is not really much.

Good luck with those NVidia chipsets... :)

Something i use to stress GPU's to the max is this program! great for me obtaining my max 24/7 stable OC..

http://downloads.guru3d.com/download.php?det=1717

Set the benchmark Res to max ur screen takes and select stabilty mode/full window mode and let run for 45-60 mins

If u get hard lock up then obiously problem there!

everest is only good for, well not really much.
its benchmark is only that, a benchmark its not a stress test.
its hardware monitoring is subjective.

go download 3dmark 06 from the futuremark website.
this will crash or give you odd artifacts or even crash while running the tests.
by 'artifacts' i mean you may see wierd things on the screen, discoloration, wierd pixels doing wierd things. textures that are missing or otherwise, not normal. checkerboard colored things on screen.

also what is your cpu and system temps.

you can download coretemp or speedfan or the tool that came with your mobo.
i prefer speedfan because it already has the core temps built into it, and it also shows video card temps, northbridge temps and any other sensors it finds.
several readings ill be wrong,. its just the nature of the program so you click configure and uncheck them so they do not show up.

30 minutes of memtest isnt enough although you will usually see errors pretty darn quick if your ram is bad enough.

i also mentioned showing or writing down the bios settings. so please can you do so. this is really really going to help us help you. when you have the time to do so of course.

i have a feeling you wont find any issues unless its directly video card related.
if you were overclocking before you probably had it way wrong thus you got the system crashes.

Quote:

---

Originally Posted by Lestat

everest is only good for, well not really much.
its benchmark is only that, a benchmark its not a stress test.
its hardware monitoring is subjective.

go download 3dmark 06 from the futuremark website.
this will crash or give you odd artifacts or even crash while running the tests.
by 'artifacts' i mean you may see wierd things on the screen, discoloration, wierd pixels doing wierd things. textures that are missing or otherwise, not normal. checkerboard colored things on screen.

also what is your cpu and system temps.

you can download coretemp or speedfan or the tool that came with your mobo.
i prefer speedfan because it already has the core temps built into it, and it also shows video card temps, northbridge temps and any other sensors it finds.
several readings ill be wrong,. its just the nature of the program so you click configure and uncheck them so they do not show up.

30 minutes of memtest isnt enough although you will usually see errors pretty darn quick if your ram is bad enough.

i also mentioned showing or writing down the bios settings. so please can you do so. this is really really going to help us help you. when you have the time to do so of course.

i have a feeling you wont find any issues unless its directly video card related.
if you were overclocking before you probably had it way wrong thus you got the system crashes.

---

Thats just the thing! I haven't OCd ANYTHING yet! The board is running, with the exception of a few CPU options, completely stock out ofthe box. All I had to change was the CPU multiplier to bring the chip back up to its stock speed of 2.933GHz.

This is a stock configuration, crashing under load (i.e. when playing games), and displaying intermittent issues. And yes, those tests werent ran as long as they "should" be, but I think so STOCK BIOS settings it should be long enough. Just have to stress test the vid cards one at a time and see what turns up...I'll report back later today with screenies of my BIOS just to show you guys what the default settings are.

--skrips

well stock bios settings arent good for all hardware, its just a general starting point.
ram being a big one.

you have the 6400 Dominators which absolutely require more voltages.
if you have not done so yet increase the ram voltage to 2.1 or 2.2 just for right now and see if things clear up.

this isnt about overclocking this is simply trying to find out what the problem is.

also take a picture of the hardware monitoring page that shows all your voltages and temps.
i have a nseaky feeling if your running the dominators and haven't touched the voltages that your ram is not getting enough.

it could also be a bad hard drive or anything truly.

here is what you do. unhook the ide cables from any cdrom or extra hard drive you have.
run 1 stick of ram
take everything out of the motherboard that doesnt need ot be there.
run only the minimal things to get you into windows.

then start doing your normal things and see if it crashes.
if it does then you know that those items you unhooked werent causing any issues.

default bios settings are usually ok but it doesnt mean they work for your hardware

i see you are also running 2 of the GTX video cards.

there is a problem in itself. OPB, a good standing and knowledgeable forum member is having similar issues and he is running dual cards. (i believe) but on an intel based chipset.
by the way he described his issue, it sounds like a video card so you could be having the same issues.

only use 1 video card. for now and see what happens.
if it still locks up the swap out to the other video card.
same for the ram run one stick, if it still happens then swap out to the other stick.

if you have an old 775 cpu laying around change the cpu, see if it still happens.

***final comment of about the pictures***
keep the resolution low. there is no need to use a ful 5mp or anything like that.
jsut set it to the lowest resolution of your camera.
and then resize them in windows to something like 1024x768.
you can let the forum host them but make sure they are jpeg pictures and less than 200k EACH. anything higher in file size and the forum will reject them

if need be i can host them for you and email you the links.

Quote:

---

Originally Posted by Omastar

Give ATI Tool v 0.26 a try. It runs an OpenGL (I think) graphics benchmark continually, and if you notice any graphical abnormalities while it's running (yellow or white spots, other discolorations/anomalies) with your cards at stock, you likely have a dying card. You could also give Rthdribl a try, I suppose. Dunno if that's really suited for stability testing.

Why not just pull one card and test each one separately by playing a game or three? It's also possible you have a bad SLI bridge, I guess.

---

Can SLI bridges really be bad? I had two of the dual-sli bridges come with the Abit board, and I have 1 tri-sli bridge that came with the Evga board (this is the one I am using now). I highly doubt all 3 of them are bad, since I have tried all three of them to verify it was not the bridge. Was one of the first things I tested :)

--skrips

Quote:

---

Originally Posted by Lestat

well stock bios settings arent good for all hardware, its just a general starting point.
ram being a big one.

you have the 6400 Dominators which absolutely require more voltages.
if you have not done so yet increase the ram voltage to 2.1 or 2.2 just for right now and see if things clear up.

this isnt about overclocking this is simply trying to find out what the problem is.

also take a picture of the hardware monitoring page that shows all your voltages and temps.
i have a nseaky feeling if your running the dominators and haven't touched the voltages that your ram is not getting enough.

it could also be a bad hard drive or anything truly.

here is what you do. unhook the ide cables from any cdrom or extra hard drive you have.
run 1 stick of ram
take everything out of the motherboard that doesnt need ot be there.
run only the minimal things to get you into windows.

then start doing your normal things and see if it crashes.
if it does then you know that those items you unhooked werent causing any issues.

default bios settings are usually ok but it doesnt mean they work for your hard drive.

i see you are also running 2 of the GTX video cards.

there is a problem in itself. OPB, a good standing and knowledgeable forum member is having similar issues and he is running dual cards. (i believe) but on an intel based chipset.
by the way he described his issue, it sounds like a video card so you could be having the same issues.

only use 1 video card. for now and see what happens.
if it still locks up the swap out to the other video card.
same for the ram run one stick, if it still happens then swap out to the other stick.

if you have an old 775 cpu laying around change the cpu, see if it still happens.

***final comment of about the pictures***
keep the resolution low. there is no need to use a ful 5mp or anything like that.
jsut set it to the lowest resolution of your camera.
and then resize them in windows to something like 1024x768.
you can let the forum host them but make sure they are jpeg pictures and less than 200k EACH. anything higher in file size and the forum will reject them

if need be i can host them for you and email you the links.

---

Definitely sounds good. Unfortunately, I dont have an extra CPU lying around. I did, however, run the machine with only 1 card. Was the first thing I did early yesterday before I started posting. The machine seemed to run fine with both cards individually, then I put them both back in and it started crashing. Then later last night it started crashing with only 1 card in the box.

Sooo, I will try upping the ram voltage, I will take BIOS pics, I will test the cards seperately, and I will get back to you guys later today.

Again, Thank you so much for the help, I really needed it apparently. Accepting help from people who know more than you is the only way to learn!

--skrips

there is one thing i need to ask

what exactly is happening. to a T i need to know.
i also need to know exactly what you are doing when this happens.

this has really got me curious.

last thing i have a blue tooth mouse and keybaord that sometimes, when it goes to sleep, the mouse and keybaord i mean (when i am away form the pc for a while) and i come over and wiggle the mouse it was blue screening my pc.
i believe after looking into it it was a fualt of my own for misconfigured voltages in the bios as it hasnt happened for a while after i redid my overclock, but its just a point im making about mice and keybaords doing wierd things.

I really havent gotten to deep into your problem but seriously I have some power supplys that have done some crazy stuff.

Once my comp wouldnt start and thought it to be mobo because psu would flash and it would measure right volts on all rails and everything... but after I tried a different psu it worked on my rig and my old one still worked for other rigs. It's crazy but true. It may not be your psu causing crashes but I would never rule it completely out.

Also flash the bios if you are running any memory above 800mhz. The new bios fixed the issue.

Are you experiencing crashes after a few hours of computer work? Maybe you got thermalgrease on the bottom of the chip (just saying what I've seen)... the grease could be heating up and shorting contacts (since it has metal shavings)

Lastly are your giving the correct amount of voltages to the memory? This is crucial even at stock timings and speed. You should be setting voltages to manual settings as per manufacturer.

Quote:

---

Originally Posted by Skrips

It hardlocks, the screen goes blank, then it comes back up with distorted and frozen graphics image...Now, since this is THE EXACT SAME THING HAPPENNING ON TWO DIFFERENT BOARDS FROM TWO DIFFERENT COMPANIES, what could it POSSIBLY be?

---

This symptom leads me towards the video card(s). I'd borrow a friends card and see if you have the same problems. Or, even better, put you card(s) in a known stable system and see if it creates problems on that other system.

Well if you are having display issues it definitely points to a vidcard. Download the free version of all the 3dmarks and just loop it continuously. Oh and serve you right for using Nvidia chipsets...I kid I kid :P

Quote:

---

Originally Posted by situman

Well if you are having display issues it definitely points to a vidcard. Download the free version of all the 3dmarks and just loop it continuously. Oh and serve you right for using Nvidia chipsets...I kid I kid :P

---

Rofl! :p:

AHHHAHAHAHAHAHA! I think I found the problem! Usuing the fur rendering video stress test found here:

http://downloads.guru3d.com/download.php?det=1717

I ran one card at a time. The first card displayed HORRIBLE graphics glitches and rendering glitches, and was running at about 1 fram every 30 seconds, and NOT even displaying the FPS or the test data. I popped the second card in, and guess what? It is running like a champ on crack and currently crunching out a nice fuzzy doughnut at an average of 35 FPS. Thanks to Psycho_eddie for the link and the suggestion to use it!

What I don't get is this: In the real world testing (World of Warcraft in this case) why was it "working" from time to time in SLI even with a bad board? Shouldn't it have not worked at all, peariod? And I swear, back when my machine was first having issues, I tested both cards individually and had issues. Now it seems like there are no issues with the card I am currently running.

Well, I will be letting it run, and see what happens. I will also try the real world testing soon, and then I will post an update. Then, after its PROVEN to be the card, I think an RAM to BFG would be in order...After that, hey, maybe I will finally be able to OC this thing with your help!

Thanks to all who posted with info, you have all been a GREAT help to me, and I am truely indepbted for your sevices. Keep watching this thread and I will keep everyone up to date!

--skrips

wow one card that bad ? seems pretty odd.

the reason the other games were running is because they were not even using SLI, they were using the onie card that runs good.
and on the occassion when it maybe did do something with the bad card it crashed.

thats how i see it.

the whole SLI thing is nothing but a joke. they have never gotten it to work properly and games and game programmers just do not care about SLI because its such an expensive thing for the average game player to have to afford.

in reality sli will increase your performance but in truth you wont see huge gains, atleast the way i see it, until you are at 1900x1200 or higher, thats where the dual cards really start to show their power.

yes you get improvements on the lower graphics but not like a higher resolution.

RMA that bad card tommorrow if you still have warranty

Quote:

---

Originally Posted by Lestat

wow one card that bad ? seems pretty odd.

the reason the other games were running is because they were not even using SLI, they were using the onie card that runs good.
and on the occassion when it maybe did do something with the bad card it crashed.

thats how i see it.

the whole SLI thing is nothing but a joke. they have never gotten it to work properly and games and game programmers just do not care about SLI because its such an expensive thing for the average game player to have to afford.

in reality sli will increase your performance but in truth you wont see huge gains, atleast the way i see it, until you are at 1900x1200 or higher, thats where the dual cards really start to show their power.

yes you get improvements on the lower graphics but not like a higher resolution.

RMA that bad card tommorrow if you still have warranty

---

Luckily BFG has a lifetime warrenty on all their cards :P

And yeah, I can't believe 1 card was bad. It just doesnt make much sense, but I agree with the way you put it. Oh well, least I didnt have to smash my machine :)

I once had a customer with a similar problem, one of his two GTX cards was faulty. Unfortunately, we only discovered that AFTER I converted his whole rig into w/c...
I saw it in 3DMark06 Canyon Flight, there were a few yellow and red dots. Tested both cards separately with ATITool and voila... so it does happen and you're not alone in the Universe :D

Good luck with your rig!

Quote:

---

Originally Posted by Skrips

I tried 3 different graphics cards...

---

So the problem is a graphics card?
:confused: ;) :)

Quote:

---

Originally Posted by Plaicd

So the problem is a graphics card?
:confused: ;) :)

---

lol

it would be if he was always running SLI..

he did say it ran fine for a while

its just that the fur rending benchmark is pretty intensive and that showed immediately the problem.

personally for the sake of argument i would uninstall all your drives. set SLI to off and run that one bad card thru some benches like 3dmark or the fur rendering and see if it still shows up as bad.

Quote:

---

Originally Posted by Lestat

lol

it would be if he was always running SLI..

he did say it ran fine for a while

its just that the fur rending benchmark is pretty intensive and that showed immediately the problem.

personally for the sake of argument i would uninstall all your drives. set SLI to off and run that one bad card thru some benches like 3dmark or the fur rendering and see if it still shows up as bad.

---

Yeah, I agree. Just for a baseline though, I am running the good card still. Been running for almost 2 hours now on the fur render, no glitches so far!

Before I RAM I will try and make sure the other card IS bad. Also, I have not always been running in SLI. It seems that anytime the bad car was plugged into the mobo, either as primary or secondary card, it would cause problems. Then running one card at a time through the fur render allowed me to spot WHICH card was the faulty one. /shrug

Now we know! :up:

--skrips

After reading this whole thread up till this post. I am still curious what voltage is being supplied to his corsair dominators.

Skrips please tell us the voltage being supplied to your memory via bios.

Quote:

---

Originally Posted by punisher69

After reading this whole thread up till this post. I am still curious what voltage is being supplied to his corsair dominators.

Skrips please tell us the voltage being supplied to your memory via bios.

---

Well, the ram was never the issue apparently...but fyi it was running at stock speeds: 5-5-5-18 2t@1.85v. According to ALL the suggestions I got, I upped the RAM voltage back up to 2.1v (which is where I had it right after I put the machine together, it was one of the first things I did). Until I test further on the ram side, I dont know for certain that the extra voltage is going to do anything. But I can tell you this: This ram is supposed to run at 4-4-4-12 2T@ 2.1v, and I am by god going to get it to run at those speeds.

Cheers! :toast:

--skrips

FYI, after my last post, I decided to up the timings back to the rated settings. I am now running at:
4-4-4-12(2t)@2.1v

I am doing a Prime95 blend run(2 instances) and I will let you guys know how it works.

--skrips

good to hear.

you probably never had an issue with anything other than the video card. but once you run a full suite of tests you will know for sure.

Quote:

---

Originally Posted by Lestat

good to hear.

you probably never had an issue with anything other than the video card. but once you run a full suite of tests you will know for sure.

---

Yeah, I've learned my lesson for sure. I am never going to make that mistake again. :up:

--skrips

This thread makes me feel warm and fuzzy inside :) Props to everyone who helped Skrips out!

Quote:

---

Originally Posted by Lestat

well i will say your first initial issue was that putrid abit board.
the 680i was not one of abits better devices. not that i ever saw, but i didnt hang around to watch it get fixed or not.

---

seems like most people are happy with BIOS 14.

Personally, I'm staying clear of nVidia chipsets on Intel for a while.