Bitcoin Forum

Yesterday I had an AMD/ATI radeon 5870 working just fine. ( never overclocked )

Today the computer will not complete POST. I have tried the card in 2 different systems:
- Award BIOS postcode stops at 26
- AMI BIOS postcode stops at 2A


The computer boots up with other video cards, just not the 5870.


Is there anything I can try to get this card working again ?

both codes point to ram!, strange but true.

asuming ami bios is pheonix then 2A - Initializes different devices through DIM.
are they paired or are you clocking the memory? If another card works in this system then i'm stumped.

Have you been overclocking? What were the temperatures like when mining? It looks like you've damaged one of the memory chips - in which case, there's probably not much you can do.

The card is working again. I think this card has to "warm up" before it will complete the POST.

On a cold boot it stopped at 2A. Seeing that it was broken I just kept rebooting it and noticed the post code got a little farther than 2A. I waited some more to heat up, rebooted and it finally booted into the OS. ( Either that or the position of one of the PCI express power connectors was under a little pressure. )

I saw posts online of people baking their gpu in the oven to fix it. Before I went that far I was going to try a few other things. Now it is back to mining.

My GPU has been running at 85 degrees.

I wanted to mine at a safe level without any overclocking. So I followed the executive summary of:

Single pool, dedicated miner:
cgminer -o http://pool:port -u username -p password -I 9

That results in gpu running at 85 degrees. The readme said all the defaults should be safe values.

I have now added --auto-fan so that the temp comes down to 75 degrees.

What is a safe temperature to mine at ?

The lower the better. I usually run mine at 75, but that's just me.

Me too. 

I'm pissed, I just lost a card yesterday too.

Noticed I checked on my mining and had no shares submitted in like 20hrs, and go to restart the machine and couldn't remote, it was hard locked so I restarted it and it had like 5 rapid beeps and the GPU fan started at 100% throttle. Powered fully down and turned off PSU, turned all on, and it starts to windows (no monitor so software-wise) and I can log in remotely since VNC uses software rendering, but as soon as I try to start mining fan jumps up and computer hard locks. :|

It was running 80-85 under normal conditions of full mining, I had fans set with a 85 ramp up as it approached up to 100% so that it kept that as a ceiling.

I find it really interesting that it's letting it POST from a cold boot until put under load. A warm restart is the only time it doesn't post after an error locked it up.

well there's your problem

Perhaps the default settings in the miners should have a lower temperature to run the hardware safely. ( 75 sounds like the recommended temp )

Then only allow the overclockers to run it hotter.


How long had you been running the card from 80-85 ?

You can overclock a 5870 and still keep it at 75 or lower.  I use MSI Afterburner (http://event.msi.com/vga/afterburner/download.htm). 

From the cgminer docs:


That maintains an 85 degree gpu.

All I am saying is that when that exact command is issued it should keep the gpu at 75 degrees.

The card worked for about 24 hours and now it stopped again.

I see people are selling GPUs as parts or repair.... how much is a 5870 worth when sold as parts/repair ?

you can go to electronics shops/services to change the gpu part

I guess they do have some value. Here is a non-working one currently bidding about $50 http://www.ebay.com/itm/HIS-ATI-Radeon-HD-5870-AS-IS-non-working-/221075979574 (http://www.ebay.com/itm/HIS-ATI-Radeon-HD-5870-AS-IS-non-working-/221075979574)