Get yourself the latest beta of Everest, there is a thread in the News section. This program is very good.Quote:
Originally Posted by Skrips
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.
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.Quote:
Originally Posted by Lestat
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.
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 :)Quote:
Originally Posted by Omastar
--skrips
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.Quote:
Originally Posted by Lestat
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.
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.Quote:
Originally Posted by Skrips
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:Quote:
Originally Posted by situman
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
Luckily BFG has a lifetime warrenty on all their cards :PQuote:
Originally Posted by Lestat
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!
So the problem is a graphics card?Quote:
Originally Posted by Skrips
:confused: ;) :)
lolQuote:
Originally Posted by Plaicd
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!Quote:
Originally Posted by Lestat
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.
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.Quote:
Originally Posted by punisher69
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.
Yeah, I've learned my lesson for sure. I am never going to make that mistake again. :up:Quote:
Originally Posted by Lestat
--skrips
This thread makes me feel warm and fuzzy inside :) Props to everyone who helped Skrips out!