Video Card Driver Issues

[RehabOC80]

Hey all, I'm having some issues with my computer(I'm assuming it's Video Card software related) but I'll explain what's going on.

Computer:
Windows 7 64-bit
Quad-Core 2.5ghz
Radeon 5670 Vid Cards(x2 w/ Crossfire)
4GB RAM

Problem:
Whenever running a graphic intesive program(WoW/certain FB game/HoMM5, etc) every so often I run into a few different possible issues. Either 1.) The Screen completes freezes and have to restart 2.) Screen goes one solid color w/ vertical lines(can be different colors, sometime Red, purple, blue, black, etc) 3.) Screen goes black for a few moments, then eventually comes back w/ an error pop-up in the corner stating "Windows has recovered from a failed Video Driver"(or something to this extent.

So, of course the first things I've done were rollback the Driver, was thinking maybe I had a corrupted updated version, and tried w/ an older driver. Didn't work. Tried updating the driver again. Didn't work. I've run some clean-Ups, Scans, thought it may be a hardware issue so opened computer up and cleaned it out/dusted it(it was pretty damn dirty) and that seemed to put it off for a bit, so it made me think it may be a hardware issue(has a liquid cooler, had the comp for 2 years and never changed the liquid in it...oops X) ) but the computer never feels warm(usually cool tbh) and never had any overheating warnings.

Anyways, any thoughts on the issue? It's happened quite a bit tonight, I might open it up and try cleaning it out again, and I haven't tried any of the fixes I mentioned in a couple months, since it first started going on.

Appreciate any insight! Lmk any questions/info you may need! I'm pretty experienced w/ computers, so hopefully I can get some decent advise.

[BluEyedDemon]

I am running almost exactly what you are, and I am getting the same problems when I run a game like WoW.

[Mordoran]

I got the driver crash when i was OCing my Geforce card. Solved it by turning up the core voltage on the card to 1075 mV.

Now you might not have OCed your cards but who knows, might still be unstable. Use a program like MSI Afterburner and turn it up to 1025 or 1050 and see if that does anything. I've heard rumours that some cards can become unstable so you might just be one of the unlucky ones.

Do note that you should keep an eye on your temperature of the GPU core if you try this. Everything below 1100 mV is usually fine with a stock cooler but just in case


Alternative would be to install an older driver (yes older :P). Had an issue with a 400 series card after I updated the drivers. Once I rolled back the driver it worked just fine.

Btw what size of PSU do you have in that system?

[RehabOC80]

I am running almost exactly what you are, and I am getting the same problems when I run a game like WoW.

I've googled the issue and didn't come up with too many things either w/ the same problem OR solutions.

Have you had any luck?

I got the driver crash when i was OCing my Geforce card. Solved it by turning up the core voltage on the card to 1075 mV.

Now you might not have OCed your cards but who knows, might still be unstable. Use a program like MSI Afterburner and turn it up to 1025 or 1050 and see if that does anything. I've heard rumours that some cards can become unstable so you might just be one of the unlucky ones.

Do note that you should keep an eye on your temperature of the GPU core if you try this. Everything below 1100 mV is usually fine with a stock cooler but just in case


Alternative would be to install an older driver (yes older :P). Had an issue with a 400 series card after I updated the drivers. Once I rolled back the driver it worked just fine.

Btw what size of PSU do you have in that system?

I haven't OC'd it, and it's not a stock cooler, it's a Liquid cooled system, so I haven't noticed any temperature increase(I don't have an LEd thermometer though, but haven't gotten any notifications of any cooling issue.

The first thing I tried was rolling back my drivers, and didn't seem to change anything for me. I believe I did a system restore back to an older point as well(not 100% sure if I did do this or not, as it was a while ago that I rolled back my drivers) but that didn't seem to fix much.

Since it dual Vid cards, could it be something w/ the xfire Tech? or if I switched the seating of the two would that change anything?

As for Power Supply it is...(Idk I'll open it up and take a look in a bit because I can't remember)

I put it together mostly myself, from IBuyPower

edit: If I can resolve this issue, I will probably OC it, I just never got around to doing it(the Comp is ~2years old now) but I'm sure it could handle quite a bit of OCing.

[Mordoran]

Well the reason I ask for your PSU is that it might be too small to feed 2 cards. You sure its big enough?

And well if you have a watercooled system then you wont get any temperature problems. If the PSU is big enough, try to increase the voltage on the GPU to 1050 and see if it becomes stable.

Edit: Come to think of it, I do recall me having this sort of issue with an ooooold ATI card (bought ATI once, never will again), and it caused me so much pain. When it started to crash my system i simply replaced it with a new card (nvidia in this case) and the problem went away.

If you are really unlucky one of your cards might be faulty. Try taking out one and run with the other and see if the problem persists.

[RehabOC80]

Well the reason I ask for your PSU is that it might be too small to feed 2 cards. You sure its big enough?

And well if you have a watercooled system then you wont get any temperature problems. If the PSU is big enough, try to increase the voltage on the GPU to 1050 and see if it becomes stable.

Edit: Come to think of it, I do recall me having this sort of issue with an ooooold ATI card (bought ATI once, never will again), and it caused me so much pain. When it started to crash my system i simply replaced it with a new card (nvidia in this case) and the problem went away.

If you are really unlucky one of your cards might be faulty. Try taking out one and run with the other and see if the problem persists.

I'll take a look at it tomorrow. But if it WAS the PSU, wouldn't it have had this problem from the get-go? This didn't start until a few months ago, so it ran for 1.5+ years without an issue.

For the Card thing, Yeah I'll give the 1 a shot, or just switching them, I'll try a few different things and see if that changes something. (Sorta unrelated, when I first got my comp 1 of my Vid Cards was fauly, and had to send it in and they sent me a new one, worked fine after that, so maybe the 2nd one is doing that now?)

If it is a Vid Card...well, Xmas is coming up @least right...lol

[Mordoran]

Well even the best PSUs can fail with time Might have been that your PSU were just barely keeping it up when it was new but time takes it toll on electronics. The transformer inside can't decline, but the electronic components that converts the power to DC can. Again its not by any means necessarily the case for you but its an option.

As for the cards, I'd suggest you try out one at the time. See if one makes the error of if it runs smoothly. If any one of them crashes you got your answer, if none does it seems like a PSU problem.

[Isendar]

Let's try a few things. Here's the steps I followed when I got BSOD + restart on my 560Ti SLi setup:

Hardware Steps:
1. Switch GPU's positions.
2. Ensure that both GPUs are firmly seated in their PCI-e slots.
3. Make sure that the X-fire/SLi bridge is firmly in place.
4. Switch the position of the X-fire/SLi bridge (This may not matter, but it did for me!)
5. Test each card by themselves

If none of the hardware steps solve your issue, try these software steps:
1. Use MSI Afterburner to check temps. If ok, move to step 2.
2. Download FurMark and utilize the burn-in test feature.
3. If burn-in test does not generate an error go back to MSI Afterburner.
4. In MSI Afterburner, downclock the gpu settings and test in game that was causing issue.

Note: Its entirely possible that out of your two cards, one is bad. I noticed on my newest 560 Ti (same brand and model), my GPU temp is always 10 degrees hotter than my older card. First card is only 2 weeks old, second is 3-4 days old. When I switch positions, I get the same results. Go figure! It is possible that your PSU can't handle the two cards. The PSU may be able to supply enough power to the GPUs when idle, but once ramped up in a game, the demand across the 12v rail can become too much for the PSU. Alot of power is drawn from both the CPU and GPU when playing games, so the possibility that your PSU doesn't fit the bill, can be likely if you barely made the minimum 12v rail requirement!

[RehabOC80]

Let's try a few things. Here's the steps I followed when I got BSOD + restart on my 560Ti SLi setup:

Hardware Steps:
1. Switch GPU's positions.
2. Ensure that both GPUs are firmly seated in their PCI-e slots.
3. Make sure that the X-fire/SLi bridge is firmly in place.
4. Switch the position of the X-fire/SLi bridge (This may not matter, but it did for me!)
5. Test each card by themselves

If none of the hardware steps solve your issue, try these software steps:
1. Use MSI Afterburner to check temps. If ok, move to step 2.
2. Download FurMark and utilize the burn-in test feature.
3. If burn-in test does not generate an error go back to MSI Afterburner.
4. In MSI Afterburner, downclock the gpu settings and test in game that was causing issue.

Note: Its entirely possible that out of your two cards, one is bad. I noticed on my newest 560 Ti (same brand and model), my GPU temp is always 10 degrees hotter than my older card. First card is only 2 weeks old, second is 3-4 days old. When I switch positions, I get the same results. Go figure! It is possible that your PSU can't handle the two cards. The PSU may be able to supply enough power to the GPUs when idle, but once ramped up in a game, the demand across the 12v rail can become too much for the PSU. Alot of power is drawn from both the CPU and GPU when playing games, so the possibility that your PSU doesn't fit the bill, can be likely if you barely made the minimum 12v rail requirement!

Since it doesn't happen everytime I play WoW, is there a way to force a strain on it and see if it will happen? so I don't have to open it, take one out...play for a while, etc etc?

[Sephiracle]

Since it doesn't happen everytime I play WoW, is there a way to force a strain on it and see if it will happen? so I don't have to open it, take one out...play for a while, etc etc?

Furmark will use a lot more power than WoW. You'll find out whether its a power issue very quickly.

[RehabOC80]

Furmark will use a lot more power than WoW. You'll find out whether its a power issue very quickly.

Which is what exactly?

Well even the best PSUs can fail with time Might have been that your PSU were just barely keeping it up when it was new but time takes it toll on electronics. The transformer inside can't decline, but the electronic components that converts the power to DC can. Again its not by any means necessarily the case for you but its an option.

As for the cards, I'd suggest you try out one at the time. See if one makes the error of if it runs smoothly. If any one of them crashes you got your answer, if none does it seems like a PSU problem.

btw Xion 700w peak PSU - lmk if you wanna know the Amperage/voltage/range/etc.


And I opened it up, swapped the seating for the Xfire, pretty dusty/dirty so tonight I'll clean it out before I reseat the GPU's, but they seem seated fine where they are.

[chaley]

The best thing to do in regard to testing would be to take each card and put them in one at a time in each slot. Run a benchmark test (the poster above is right about Furmark) and if you have issues try to isolate what happened. Is it the port on the motherboard? Is it the card? Is it xFire? Once you've done 2 tests for each card (1 test per slot), try both together and swap them around. This kind of process can take several hours to test, but without testing equipment like what I have (very nice, but expensive stuff) it's the best way to see what the problem is. If you can get your hands on a PSU tester (usually $20) that might shed some light on the issue as well.

---------- Post added 2011-11-16 at 01:00 PM ----------

This software is what I sometimes use to do benchmarks.

[RehabOC80]

The best thing to do in regard to testing would be to take each card and put them in one at a time in each slot. Run a benchmark test (the poster above is right about Furmark) and if you have issues try to isolate what happened. Is it the port on the motherboard? Is it the card? Is it xFire? Once you've done 2 tests for each card (1 test per slot), try both together and swap them around. This kind of process can take several hours to test, but without testing equipment like what I have (very nice, but expensive stuff) it's the best way to see what the problem is. If you can get your hands on a PSU tester (usually $20) that might shed some light on the issue as well.

Awesome program! Thanks! (although, not to be used by those that have possible seizures :P)



---------- Post added 2011-11-16 at 01:00 PM ----------

This software is what I sometimes use to do benchmarks.

Anyways, my overall score was 1331.8(which seems pretty high actually) but I'm not exactly sure what I should be comparing it to. Any tips you can help me with what I should be doing to diagnose? (e.g. should I run it multiple times w/o 1 of the cards, etc) and what would I be looking for?

---------- Post added 2011-11-16 at 11:37 PM ----------

So far reseating the GPU hasn't done anything, next I'll try to swap em nd run the tests. What am I looking for in the Tests? and should I just do the GPU tests instead of a full benchmark each time?

[RehabOC80]

Still not sure how to compare the tests, there will obviously be a drop if I remove one, but how will I know if one is a bit faulty?