Random lock-ups

[Canowyrms]

As the title says, my computer has been randomly locking up since last night. I have a hunch that my graphics cards are behind this. I've been toying with mining (bitcoin) lately, and haven't run into any sort of problems. For those who are unfamiliar with mining, it utilizes the GPU's ability to perform repetitive calculations. When mining, the GPUs are both under 99-100% load, and with that being said, they tend to get quite hot. To help manage the temperature, I downloaded MSI Afterburner to underclock the GPU memory clock. My cards (system specs will be below) come with a stock memory clock of 1200 MHz, and I clocked them down to 600 MHz. It stayed stable for a few minutes, then all I got was a white screen. No HDD activity, no control+alt+delete, nothing. Just a hard reboot with the button on my case. I rebooted, thought it was weird, and went on with my day only to have the same thing happen. I've been toying with the configuration and it seems like I cannot rid myself of the problem. The screen has gone white, black, reddish-white with vertical bars.

I have uninstalled MSI Afterburner since I reset the memory clocks. I have uninstalled/reinstalled the AMD Catalyst Control Center (drivers and the whole nine yards). I still get the problem.

Here is where I think it gets weird. I have two graphics cards, and have tested each card on its own. Both seem perfectly fine (system is stable, no lockups). I have tried running both cards together using different PCI-E slots than in the original set up, and they're fine. When I revert back to the original setup, I experience more lock ups. I also have tried using alternate CFX cables with no avail.

Now, I have taken a peek into the event viewer, but I'm not exactly sure what I would be looking for, and it seems like whatever I find is irrelevant to this issue. I have the cards running fine right now in an alternate configuration, but this configuration doesn't allow me to have my audio card plugged in, so I will not keep this configuration.

Please let me know if I can provide any additional information.

System specs:
ASUS M3A32-MVP Deluxe WiFi/AP edition
AMD Phenom x4 9950 BE 2.6GHz
4GB Mushkin Ram (I can look up the exact specs if required)
Antec High Current Gamer 620W Continuous (Newer than other parts, never had a problem with it)
2x HIS Radeon HD 6770 850MHz Core 1200Mhz Memory
HT Omega Striker 7.1

 

[QuietPC]

So if you run one card only everything is fine?

My experience is that running dual cards almost always leads to challenges so I run the best single I can afford. If one card does the job then just run one. There may be a software issue or conflict that simply isn't apparent. Consider selling your two 6770's and get a single 6870.

 

[Canowyrms]

I run two 6770's because together, their performance matches that of the best single 6xxx series card. I haven't had any issues up until last night *shrugs*

I would rather keep what I have and work out a solution that can put me back in the original set up I had than get a new card.

 

[QuietPC]

Then its either a software issue (driver and/or game(s)) or you damaged one or both during intense use and heat. Are they still under warranty?

 

[daship]

Mining BitCoins have been around for a long time, the days of profiting from it are far over. Especially with one PC. If you join a huge pool, you get even less coins, making it even less profitable. You will never make enough money to pay for wear/tear on hardware, not to mention elictricity bills.

My guess would be you stressed the cards to hard and ruined one. Or maybe you stressed the PSU.

 

[Canowyrms]

@QuietPC - They would average less than 70 Celcius even under 100% burn. They are still rather new, less than 1 year old for sure. Yes they are under warranty; if I can prove that the graphics cards are indeed the issue, I can have them replaced on the spot. but if each card runs fine, and the only time I have problems is during the original layout...

Also, I wasn't playing any game when this occurred. I have been on my desktop, and would randomly experience these lock-ups. sometimes when I open event viewer, sometimes when I open CCC.. I don't know what it is for sure. I will try uninstalling the driver suite again completely, and will try to clean out the registry (it probably needs to get done anyway).

@daship - I know it is not profitable - I'm not mining for a profit. I simply do not want to exchange my money for bitcoin. and besides, AMD cards have the advantage over Nvidia cards. You can technically exceed breaking even with AMD cards, but the numbers are certainly nothing to be excited about.

With that being said, I will report back when I have any other news (Will be AFK for a few hours). Thank you all for your input so far.

 

[daship]

I have bitcoined before, with your setup you wouldnt even get close to paying for 1 days worth of electricity after coining for a month.

 

[Canowyrms]

@daship - as I said, I'm not mining for a profit.

Anyway.
I've got CCC reinstalled, registry cleaned out and defragmented, and everything seems to be running fine so far.
Now, the card that my monitor is connected to will max out at 400 MHz core clock, and 900 MHz memory clock, even under 100% burn. Any ideas?

 

[QuietPC]

@daship - as I said, I'm not mining for a profit.

Anyway.
I've got CCC reinstalled, registry cleaned out and defragmented, and everything seems to be running fine so far.
Now, the card that my monitor is connected to will max out at 400 MHz core clock, and 900 MHz memory clock, even under 100% burn. Any ideas?

Any ideas for what?

 

[Canowyrms]

For my card maxing out at 400MHz and 900MHz for core and memory clock respectively. The cards come with an 850MHz core and 1200MHz memory clock stock.

During idle periods, the cards will underclock automatically, but normally go back to their stock clocks during use. Well even during 100% burn they are still stuck well below the stock point. They will remain at stock for a few minutes after boot up, but will become stuck after that.