Topic: HD5770 Problems - RESOLVED! Use Catalyst 12.11 Beta Drivers not 13.1 (Read 4597 times)

So far so good after 24 hours of continuous mining

I manually set the fan speed for 65%, which is keeping it hovering around 55C, so I think the heat that was being generated from the burn-in test was probably normal for this card under load. It seems AMD wants it to run hot on auto, probably to sell more cards after they burn out faster over time haha. I will keep in mind keeping it lower than 90%. Thanks.

Interesting.
I use 13.1 with cgminer on windows 8 and have no issues with my 5770.

Also, 81*C is not too hot for a video card on load. Although the long-term damage will be greater with greater heat, this is still cool. I personally start to worry as I approach 90*C. The cards usually throttle to protect themselves around 100+*C. A lot of 4870s that regularly ran over 90*C experienced electromigration issues a few years later.
I recommend GPU-Z to monitor your temps. I have to watch my 6950 doesn't go above 102.5*C or it will start to throttle. I wouldn't recommend that temperature for anyone. The only reason I am fine with it is because of XFX's warranty. I've already had to get them to replace the fan once.  So another bit of advice is to not run your fan regularly above 80-90% to avoid the heavy wear causing it to break.

I fully removed the latest Catalyst drivers that I tried initially, and tried the 12.11 Beta drivers as suggested above. Shares are now being Accepted! I will keep an eye on my fan and adjust speed as needed to keep it well below 80 Thank you!

I ordered a BFL Jalepeno the other day, so this is only a temporary solution until my unit arrives, but I wanted to make sure I exhausted all my options trying the 5770 due to the fact that the Nvidia card isn't pushing much.

I have my machine on most of the day anyway. I will have to do some tests with my Killawatt to see how much it's actually using when CGMiner is running.

Oddly enough, I was just playing Minecraft and getting the same MH/s as when I am not playing on my 560 Ti 

Too much heat will cause major issues.  81C is not hot enough to cause any issues with my HD6870 or 5870.
It took 104C to crash the 6870 and it still mines fine now. I was trying to cook it but it would not die so easily.
Right now I am at 78C and the fan is at 1330 RPM's.

I have my finger on the 6870 fan now......... its 85c and rising.......... 90C and still going........ 100C and I let the fan go...  back to 78C after a minute. Its right around 104C that it crashes on me

Remove and reinstall the AMD drivers without the Nvidia card installed before going nuts.  I had a similar issue that would only occur after using remote desktop to check the system. I have to reboot that PC everytime you remote in and I cant be bothered to look into it.

The Nvidia card is probably just wasting electricity mining BTC and at best it will make like .009 BTC a day(about $0.46 USD at posting). Its likely costing you a few cents ($0.04 or $0.05 USD) over the .009 BTC a day to run it and it will just keep digging you into a hole very very very slowly but surely.

If power was your issue it would likely manifest itself when stress testing the GPU but not always.

Some cards have cheap fans and heatsinks that are not suited to mining/overclocking.

I don't recommend cards with ring coolers as they don't have a large surface area.

It could just be driver related. Try a different catalyst version. I like 12.11beta.

At 81c I'd expect the fan to be higher than 62%. 90c+ is where EU lead free solder starts to get brittle.

NASA did some solder testing for Mars rovers from -120c to 85c and managed to break the solder in as little as 638 thermal cycles.  

Well, the Fan was only running at 62% during the burn-in. Would the temperature alone cause an error like that? I could try turning up the fan speed I guess.

81c is pretty hot. Bitcoin probably runs hotter than your benchmark.

Make sure there is no dust in the heatsink.

You can turn the memory clock down to save a bit of heat.

I always recommend 5830s rather than 6770/5770 they run much cooler and faster for the same price.

Hi,

I'm new to all this still, 1 week in, so bear with me please. I currently use a 560 Ti on my main system, and it gets like 88Mh/s avg. I was doing some reading and apparently AMD cards are much faster at doing the calculations. I had a HD5770 that I wasn't using, so I threw it into an older machine I use for a server now and again and setup a fresh bitcoin miner machine on Windows 7. Everything seemed to be going great until I tried to run any of the mining programs.

Initially I started with CgMiner. It started up fine and was mining away at 185mh/s, but then instead of Accepted, it said, "Verification Failed, Check Hardware". I did some reading on the cause for that error, and realized that I might not be giving the card enough power to do what it needed to do for this. So, I swapped out the 560ti on my main machine for the 5770 as it has an 1100W power supply and has the juice to power it. Same issue though... In fact, after a few attempts, Cgminer would just crash with an exception code c0000005 error. I then proceeded to try both the GuiMiner and Bitminter miners I had working on my Nvidia card, and got the same crashing issue.

I did a bit more digging and read that the 5770 might be bad somehow, so I tested it with FurMark, as suggested. I scored a 1600 with an 81C max temp at 1920x1080 on a 15min burn-in. I watched the whole 15 min test carefully and did not see any artifacts.

I have read everywhere that using the latest driver is best, so I was using the Catalyst 13.1 driver on the system.

Any ideas?

Thanks!

RESOLVED! - Fully removed Catalyst 13.1 driver and re-installed Catalyst 12.11 Beta. Watching temps and will adjust as needed to avoid overheating.