Bios flash gone wrong? crazy artifacts

[timaishu]

I flashed the bios of my asus 5850 directcu to the TOP version a few days ago. I backed up the old one using GPUz and using a guide I found, flashed mine to a TOP. The flash said to be successful. Computer restarted fine. I unistalled my gpu drivers and then wiped the remaining files with driver fusion. Then reinstalled them. MSI afterburner confirmed that it flashed because my stock core/memory were raised to the expected levels. I ran kombuster and was getting long sharp looking artifacts. Not the usual little cubes you see. Odd.... I figured something went wrong. So I re-flashed it back to my old bios. This time though I didnt reinstall graphics drivers(maybe the problem?) as my resolution wasnt all screwed up.

Anyways, it still does the same thing. My computer operates fine otherwise, just certain games are all jacked up now. Its not overheating either, and it is currently at stock settings. I took a video of it. I doubt I broke the card because even though my card isnt a reference, I flashed it with a non reference/same brand/same model bio's, just with factory overclocks set.

Can anyone think of what to do next?

Here is the guide I followed to flash it.

http://www.overclock.net/t/671675/how-to-flash-your-5850-bios

Here is the youtube video showcasing the glitching.
http://www.youtube.com/watch?v=MDhMSrkRc-A

 

[w1zz4]

Looks likes video memory problem to me.
Bios you used was probably configured for different RAM chip, using more volts (which is normally the case for OC card), so you fried yours...
Sorry to tell you that, but flashing a vid card with a different card BIOS is not a brilliant idea unless you sure component are the same... Even then... Risky
If you want permanent overclock with BIOS, you change settings in your current BIOS (dump ROM, make change, then flash back)....

 

[timaishu]

Hmmm. If I fried it though, would that mean it wouldn't work at all?

I reinstalled my graphics drivers and it didn't fix it btw.

 

[w1zz4]

Fried = physical dommage. There is no way to go back to previous state.
You may try to underclock you GPU RAM, it may fix the artifact issue.

 

[timaishu]

Any other opinions? I read through the entire thread about flashing this thing. They did mention that you should NOT flash a non reference card(mine) with a reference bios, because it will brick the card. I flashed it with a non reference bios.

I never saw anything about damaging memory.. Fuark man. Im gonna re-flash it again with my stock and see what happens..

 

[darkspartenwarrior]

Hmmm. If I fried it though, would that mean it wouldn't work at all?

Not necessarily, it's like how RAM can work but can cause erratic bluescreens and lockups, same thing with VRAM.

 

[timaishu]

Crap. So I re-flashed it again with my stock bios and still its jacked up..

Ugh, damnit...

I cannot find a single thread on this happening. Is there any program that I can bench my vram to see if I screwed it up?

 

[darkspartenwarrior]

If you open the case and look at the bottom of the memory chips on the back of the PCB you may see burn marks, if not you might have to take off the heatsink and look for damage.

 

[timaishu]

I'll take a look shortly..

I ran occt and I was getting 150,000 errors every 5 seconds or so.

EDIT: I dont see any burned marks on the bottom of the pcb. Have not take the heat sink off yet though.

 

[darkspartenwarrior]

I'll take a look shortly..

I ran occt and I was getting 150,000 errors every 5 seconds or so.

EDIT: I dont see any burned marks on the bottom of the pcb. Have not take the heat sink off yet though.

Then the card is probably just dead, the 150,000 errors are like memtest errors, they are the memory not working correctly due to damage.

 

[timaishu]

Took the heatsink off and saw no burn marks.

Man I am quite pissed. Luckily I have two 8800gts 640mb's laying around I can sli with. But still, 300 bucks down the drain.

 

[smorizio]

have you tried clocking down the core cpu and memory speed of the card?? on memory errors did you check to see if they were ram or gpu??? did you reset your system ram or remove 1/2 it and rerun the tests. also did you try the card in another pc..may not be the card can be unstable power supply. 12v rail. video card picking up ripple.

 

[timaishu]

have you tried clocking down the core cpu and memory speed of the card?? on memory errors did you check to see if they were ram or gpu??? did you reset your system ram or remove 1/2 it and rerun the tests. also did you try the card in another pc..may not be the card can be unstable power supply. 12v rail. video card picking up ripple.

It appears to be core related. With OCCT, I managed to get the errors to stop for a little once I downclocked to 640mhz. As I dropped the core speed, the number of errors continued to decrease. I ran it for a minute with 0 errors at 640core 1000memory, the suddenly it would jump to 30k. When I left the core at stock, and lowered the memory to 600mhz, the number of errors did not change(increased at a rate of 100k+).

It is being very randomly though. Testing at 600mhz and within 30 seconds it jumped to 43k errors. Then it adds maybe 5 errors every 10 or so seconds. Then it randomly jumps another 40k.

Lowered core and memory to 600mhz and at the one minute mark its at 57k errors.

The errors were GPU error checking. I have no reason to believe it is system ram related. This is getting so weird.

 

[timaishu]

Bump.. Still getting errors even at 500 core.

 

[timaishu]

I took another video..

http://www.youtube.com/watch?v=n3CpJ5LMckA

 

[Dnaangel9]

I had this same problem with my Gtx 660ti from Asus. It was the Asus DirectCU II GTX 660Ti OC edition. It took me a month to track down the reasoning's behind PC constantly crashing in games etc. Low and behold it was the Video card. The stock OC was not stable(Due to defective memory). I Too would get massive amounts of errors in the OCCT mem tests. To remedy the solution I had to lower the Memory frequency -450mhz via MSi Afterburner in order for it to run games. Now instead of crashing in games it only crashes exiting a game SOMETIMES.

With my current settings MSi Afterburner is set to:

123% Power Limit
+75mhz Core Clock
-450mhz Memory Clock
Fan: Auto

This gives me:
Core: 1163Mhz Max Auto Boost
Memory: 2554mhz

Its still BS that I have to lower a factory OC that is supposed to ship stable. I can barely even OC the core clock +75mhz over factory OC before instability kicks in. Alot of users are reporting 1250+Mhz core clocks with ease..

Lowering the memory wasnt that big of a deal. Only games it affects are memory intensive games which killed about 15-25 FPS. Not a big deal when most of my games are in the 70-120FPS range with the lowered memory clocks.

Your BIOS flash didnt mess up your card. Its that it irritated an already defective Vram that made its presence known with the new settings that tripped the instability. Atleast you got to enjoy your card, Mine was crashing games right out of that box! Love all of Asus's products from ROG Motherboards to their LED monitors..but this MAY be my first and last ASUS video card! Looks like I'll be going back to MSi for nvidia based Video Cards.


UPDATE: Would also like to point out that I also get artifacts on MSi Kumbustor for me its the K-Mark (PhysX) test. It starts out very green and blurry with vertical lines flashing on screen..after 2-3 mins screen clears up. I get no artifacts etc in games. They run great, my problem was the crashing B.S. Which I have fixed at the expense of some performance.