Weird SGminer problem

[ehalvers]

http://imgur.com/iSxFZaW

Anyone ever have this problem?

Win 7 x64

Rig is comprised of 4 Gigabyte 280x's (1 Hynix RAM 3 Elpidas) and 1 Gigabyte 270x. (pulled the one that I thought was the problem out and problem perists)

Mobo is Asrock H81 PRO BTC

8 GB of RAM

750w PSU for 1 280x and 1 270x

1200w PSU for mobo/cpu/HDD and remaining 3 280x's

Everything was running fine (one 280x was giving me problems so I had to underclock it to get the rig stable) and I had just received my newly RMA'd 1000w PSU. So I was like great I can take the 750w out and replace it with my new 1000w so I can run them all at the same clocks.

So I switched the Power Supplies and that is when problems began to arise. I start the rig up with them all clocked the same but it still seems that one card just doesn't want to run at 1060 Engineclock I have set, and crashes SGminer after an hour or so of running. So once again I began slowly underclocking the problem card to see if I can get it stable. No luck.

Then as I am tweaking the settings another problem arises. SGminer starts behaving weird and sometimes just locks the computer up completely. I restart it several times and this continues to happen. Each time I start SGminer it hangs with nothing showing or I get what s depicted in the screenshot above.

I am beginning to think I have fried the motherboard because I have tried installing a fresh OS and fresh drivers to no avail. I have also tried different versions of SGminer and they all seem to have the same problem. Once of the 280x's (always one of them) bugs out and the temp says 511 degress (always that number) and the computer completely locks up. Screenshot was taken with my phone so sorry for the bad quality.

ANY HELP IS APPRECIATED WOULD LOVE TO GET THIS RIG RUNNING ASAP. down about 3.4 MH/s

Thanks! Let me know if you need anymore info. I may have left some things out 

[Starscream]

I am beginning to think I have fried the motherboard

How could you have possibly fried your MB if the computer goes through post and loads up?? (rhetorical question).

Anyway, post full settings here because it seems you are trying to mine from wafflepool - which is scrypt using nfactor algo, so you might have conflict in settings.

[ehalvers]

I am beginning to think I have fried the motherboard

How could you have possibly fried your MB if the computer goes through post and loads up?? (rhetorical question).

Anyway, post full settings here because it seems you are trying to mine from wafflepool - which is scrypt using nfactor algo, so you might have conflict in settings.

Yeah I guess you are right haha. Just a really weird problem. worked perfectly fine for a few days before I swapped the PSU's >.<

sgminer.exe -o wafflepool etc.. -u xxx -p x --kernel zuikkis (not used when on sgminer 4.0.0) -I 13,13,13,13,20 -w 256,256,256,256,256 --gpu-engine 1060,1060,1115,1060,1135 --gpu-memclock 1500,1500,1500,1500,1500 --thread-concurrency 8192,8192,8192,8192,21568 -g 2,2,2,2,1 --lookup-gap 2,2,2,2,2

When it was working one of the 280x's was running at 1040/1450 ~650 kh/s while the rest were at 735 (elpida) 775 (hynix) kh/s.

Thanks! hope this helps solve the problem  Going to get another HDD tomorrow as I have a spare computer that only needs that to run so I can try to isolate the problem card....if there is one.

[Starscream]

Which card is it the one that doesn't wanna use 1060 core? Is it the Hynix one? Because normally Hynix and Elpida ones will use different clock speeds to get optimal speed.

BTW, have you tried running your rig without the 270x?

[ehalvers]

Which card is it the one that doesn't wanna use 1060 core? Is it the Hynix one? Because normally Hynix and Elpida ones will use different clock speeds to get optimal speed.

BTW, have you tried running your rig without the 270x?

Yeah all the Elpida RAM's are flashed with the f60 BIOS which has better RAM timing settings which allows them to reach ~735 kh/s.

The Hynix RAM can be overclocked to 1115/1500 and reach ~775 kh/s no problem (no BIOS flash either)

And the 270x is the one connected to the monitor so no I have not tried to run it without the 270x.

Connecting a 280x with an HDMI usually results in a loss of about ~100 kh/s or so as I am sure you are familiar with, while the 270x seems to only lose around 25-50 kh/s when connected to the monitor.

SIDENOTE: I have two other rigs running

1. 2x Gigabyte 7950 (using DVI plug on one of these for monitoring), 1x Gigabyte 270x, 1x Gigabyte 280x.

2. 4x Gigabyte 270's solid as a rock.

[Starscream]

Is it the same r9 280x that causes problems (511 degree)? Try swapping the slots of your cards and see if it's the card or the PCI-e slot that causes the issue.

[ehalvers]

Is it the same r9 280x that causes problems (511 degree)? Try swapping the slots of your cards and see if it's the card or the PCI-e slot that causes the issue.

I am slowly beginning to think it may be a burned out 1x PCI-e slot. It is not the same 280x that is causing problems, changes every time I try to switch it around.

I will do some more experimenting and see what happens.

Thanks again for helping me out.

Process of elimination will probably help me solve this if it is a PCI slot problem. 

[ehalvers]

This may be solved 

[ehalvers]

Awesome, fixed for the most part. Turns out one of the PCI slots decided to give out on me for some reason  Hope that doesn't happen again lol.

And guess what, another problem!!1 Mining is wonderful isn't it? 

Now one of the 280x's is slowly accruing HW errors. Messing with the settings now but nothing seems to be working.

Causes sgminer to crash after an hour or so. UGH

Thanks for the help Starscream!

[Starscream]

Awesome, fixed for the most part. Turns out one of the PCI slots decided to give out on me for some reason  Hope that doesn't happen again lol.

And guess what, another problem!!1 Mining is wonderful isn't it? 

Now one of the 280x's is slowly accruing HW errors. Messing with the settings now but nothing seems to be working.

Causes sgminer to crash after an hour or so. UGH

Thanks for the help Starscream!

You're welcome.

Which of the 280x is getting HW errors? One of the Elpida ones or the Hynix?

[ehalvers]

One of the elpidas.

Got them all running now;

Hynix 1100/1500 ~772 kh/s

2x Elpida 1060/1500 ~744 kh/s

1x Elpida 1030/1450 ~710 kh/s (Been running a few hours stable on these clocks, I think it may just be a lower quality card)

270x isn't having any problems either.

Running these on the zuikkis kernel, seems to be the most popular/efficient.