Bitcoin Forum

I'm having some problems with my Gigabyte 5970 card. The card is running at stock clock settings (also with underclocked memory). Catalyst 11.5 drivers. GUIminer with "-v -w 128 -f 10" settings.

When I start mining with that system, it runs for anything between 10 minutes and ~24 hours before crashing/rebooting for no reason at all, typically falling in the 8-12 hour range. I have been trying to monitor the temperature of the cores, and they appear to be within the normal operating range for this kind of card. No higher than 85C now that I'm running it as a stand alone card.

Anyone have any insight as to what might be wrong? Is my card broken, or can this be worked around?

If the whole system is crashing look at the minidump file ;) When I set up my mining rig I had some unstable energy conserving software that would boot up when it went idle for long periods of time and would cause the computer to crash.

Upgrade to 11.6 driver (delete all previous installs first).  Bring memory clock back to stock and see if that stops crashing (then you can start lowering it incrementally).

Also use modified phatk kernel here:  http://forum.bitcoin.org/index.php?topic=25860.0

phatk is optimized for 5870 and the 5970 is two 5870 gpus.

flags: -k phatk VECTORS BFI_INT FASTLOOP=false AGGRESSION=11

At 835oc (which is very reasonable) you should see around 380mh/s.  Even at stock clock it will blow away whatever mh/s you got with poclbm

I tried to read the minidump files with some viewer I found. I opened 5 different dumps, and all of them had a different error listed as the reason for the crash. I guess that is a dead end. Or maybe I just don't know what I'm supposed to be looking for in those dumps.

I have uninstalled the 11.5 drivers and installed 11.6 now. Going to try running phoenix with the modified kernel and see if it works better this time. Going to leave it at stock settings and see if it still runs tomorrow morning. :)

try googling what it thinks might be causing the crashes. This might lead you in the right direction as to what it could be.

Didn't work. :( Crashed some time during the night. Temperatures after around 8 hours of operation was about 76C, so it should not be the heat that is causing this. Guess I'll start googling error messages from the dumps. :)

A couple of error messages I have gotten (last 10 crashes) and the corresponding driver/file that is supposed to be causing it:


I'm not really sure what to make out of this. I seem to be having all sorts of errors, in all sorts of drivers. After spending some time googling theese errors I did notice a lot of people suggesting that several of theese could be caused by RAM issues. I have the cheapest RAM I could find installed in this particular system, so this could be the problem, but why would it run perfectly fine with say a 5850 card in it, but keep crashing as soon as I install the 5970? Any ideas?

try reseating the ram maybe you accidentally knocked it out slightly while installing the 5970?

I just ran memtest for about 24 hours, and it appers to have failed miserably. After the first 8 hours or so I had 1 error in the "red" part of the display. Now, after the temperature in the room has risen a bit, I have the whole red part filled with errors just from one test from one pass. Not sure if I can scroll this text in any way, but I'm guessing it's pretty bad.

The summary part of the memtest says I have had over 1200 errors in 26 passes.

I'm no expert on reading memtest results, but I guess it's safe to assume that this is bad. I'll have to pick up some new RAM to test if the system will run fine with the 5970.

That's too bad man it happens though. If you haven't try reseating the ram maybe it will make it better? I cant guarantee anything though.