I can't get my rig running with more then 4 GPU's

[sergio-faucetbitcoin]

Hi everyone !
I have a mining rig with 4 GPU's (Nvidia GTX 1060 6GB ZOTAC) , with a MSI 170 A-pro motherboard, and a 1000 Watt PSU.
Everything was working fine, but I decided to add 2 new GPU's. The 2 new ones are Nvidia GTX 1060 6GB AMP Edition ZOTAC.
It's nearly the same model. But since I added the 2 GPU's, all are working, I can see the fans turning on the 6 GPU's but when I start mining , I have only 3 GPU that are mining !
Only GPU0 GPU1 GPU2

Anybody have an idea of what I can do ?
Thank you !

[1715550990]

1715550990

[1715550990]

1715550990

[1715550990]

1715550990

[1715550990]

1715550990

[1715550990]

1715550990

[Supercoiner111]

Verify these have been recongised by the computer - go to device manager to check you have 6 GPU listed...

Generally better to add one GPU at a time to get it recongised on the system then shutdown and add a new card.

[Cereberus]

Verify these have been recongised by the computer - go to device manager to check you have 6 GPU listed...

Generally better to add one GPU at a time to get it recongised on the system then shutdown and add a new card.

It is always better to try adding cards 1 by 1. If this doesn't work then I think you can try to uninstall the gpu driver completely with DDU (display driver uninstaller) a program you find easily in google. Restart your PC after removing completely the driver and once you are logged in with a Microsoft Basic Display adapter as a driver, you should download latest driver of Nvidia and install it, it will normally recognize all the GPU-s this way. In my country there are a lot of persons using MSI Z 170 SLI board and none are reporting problems so far, many of them have Nvidia cards too.

[walnut]

Check your BIOS and see if it recognizes the GPUs there, if it does use DDU to uninstall your graphics drivers and reinstall them

[koody]

Check your risers, maybe bad contact.

[vectisitch]

you will have to check each card individually in the direct pcie slot on the board. once you see it's working then check a pcie riser cablle on that slot and that card. do this for each card. if all cards work then check each riser. it's most probably a bad riser cable

[sergio-faucetbitcoin]

Thank you for all your answers !

I just uninstall the drivers, and install the latest from Nvidia ! I tried all the riser, and all the cards , everything is working !
I can have 2 new AMP editions with 2 older working well together, but when I have 5 or 6 at the same time, there is always something wrong ! Sometimes I don't even have screen, it stays black screen when I start :/

I try to search on internet and tried out a few things but I can't make have all 6 running at same time ..
I uninstall the drivers again, and add the gpu's one by one. Until the 4th everything is okay, when I add the 5th I get black screen when I boot the computer .

If you have any other idea , let me know !

Thanks a lot

[Set Ready Go]

Also what OS are you running them on ?
Windows 7 dont like more then 4 cards.

Another thing you can try is to swap the gpus between your old cards and new ones so they get diffrent risers, just to verify if its riser problem or what it can be.

Good luck troubleshooting!

[Jaerin]

There is usually an option in the BIOS that allows for above 4G addressing which opens up more PCI-E channels for the cards.  I don't see that for this motherboard so you may not be able to run more than 4 cards on it.

[sergio-faucetbitcoin]

I got the option to enable more then 4 GPU's on the bios and I already did it , but after doing it, only with 4 GPU's just to try , I get a 0xc0000012f error, that tells me that :
EthDcrMiner64.exe
 "C:\Program Files\NVIDIA Corporation\NVSMI\nvml.dll" is not made to run with windows. I got the 385.41 (latest) version of Nvidia Drivers for GTX 1060 , for windows 10 64 bit. And my system is running under windows 10 64 bit.

I'm searching to resolve that now ..

Thank you

[vectisitch]

i don't like using more than 4 gpus per rig. i know people are doing it but for me it just seems to unstable. with 4 cards per rig it also makes for easier troubleshooting

[walnut]

Getting a black screen on boot is usually caused by an outdated BIOS, you can try flashing your BIOS and see if that fixes it.

[GaryH]

See if this helps, supposedly a 6 GPU walkthrough for your mobo:
https://www.youtube.com/watch?v=sa3KNS1ywzE

Go to 30:30 of the video:   MSI Z170-A Pro Walk-through for 6GPU 30:30

[taxmanmt5]

Try eliminating the original 4, and in making sure that you're getting a full driver download for the other two that you're at it. You're probably giving the board too much to think about as far as drivers are concerned. Try giving it a chance to just catch up completely with the two new cards and then come back to try to add them all together.

[sergio-faucetbitcoin]

Thank you for all the answers !  

I finally get it running with the 6 GPU's after 2 days of trying !!
I'm not really sure about what exactly made it work well  .. but it's working !

If it can help somebody, that's what I did :

I took out all my GPU's, reset my BIOS to default settings, uninstall the drivers. Then I use a VGA cable to my monitor, go to the BIOS and change the option that allows more then 4 GPU's on my Mobo (4G), and install the drivers for my graphic cards (still with no graphic card plugged in).

I added then one graphic card, turn on my computer, let him load it, and do one more restart before lauching Ethminer and look if everything is okay. I did that for each graphic card one by one, and it finally worked well !

I also overclocked them as I had them before, but I didn't do that for each card individually.
I got one GPU at 62-63° and one at 72-73° , that's a big difference, and they having the same hashrate !
I didn't push them to much, I'm dualmining around 21,5Mh/s ETH and around 210 Mh/s SIA .


I was wondering, should I overclock them individually ? And can I do that with MSI afterburner?

And another question I had, I'm mining with Nanopool , and they are telling me I get 0,7 ETH per month with 139 Mh/s average in the last 6 hours. But if I look for example on Cyrptocompare they are saying 0,99 ETH per month with 139Mh/s !! This makes a big difference !
I know the difficulty had a jump just a few days ago, so maybe Cryptocompare didn't refresh the difficulty yet ?
Or Nanopool are paying 30% less then the real rate ? If that's the case, I should move to another pool !

I also got one GPU that's makes much more noise then the others, and I can also feel it's having much stronger vibrations then the other ones when I touch it. Should I send it back to get a new one ?

Thank you !!

[Leass]

Thank you for all the answers !  

I finally get it running with the 6 GPU's after 2 days of trying !!
I'm not really sure about what exactly made it work well  .. but it's working !

If it can help somebody, that's what I did :

I took out all my GPU's, reset my BIOS to default settings, uninstall the drivers. Then I use a VGA cable to my monitor, go to the BIOS and change the option that allows more then 4 GPU's on my Mobo (4G), and install the drivers for my graphic cards (still with no graphic card plugged in).

I added then one graphic card, turn on my computer, let him load it, and do one more restart before lauching Ethminer and look if everything is okay. I did that for each graphic card one by one, and it finally worked well !

I also overclocked them as I had them before, but I didn't do that for each card individually.
I got one GPU at 62-63° and one at 72-73° , that's a big difference, and they having the same hashrate !
I didn't push them to much, I'm dualmining around 21,5Mh/s ETH and around 210 Mh/s SIA .


I was wondering, should I overclock them individually ? And can I do that with MSI afterburner?

And another question I had, I'm mining with Nanopool , and they are telling me I get 0,7 ETH per month with 139 Mh/s average in the last 6 hours. But if I look for example on Cyrptocompare they are saying 0,99 ETH per month with 139Mh/s !! This makes a big difference !
I know the difficulty had a jump just a few days ago, so maybe Cryptocompare didn't refresh the difficulty yet ?
Or Nanopool are paying 30% less then the real rate ? If that's the case, I should move to another pool !

I also got one GPU that's makes much more noise then the others, and I can also feel it's having much stronger vibrations then the other ones when I touch it. Should I send it back to get a new one ?

Thank you !!

Calculations made on pool's side are often pool round and 24h FOUND blocks diff calculations. It means that with current pool luck and with avg. 24h diff You'll get XX reward. It fluctuates often, sometimes from round to round, often during rounds.
The 10 Celcius difference problem could be caused by that card position (it could suck hot air from other cards to 'cool' itself) or by simply being primary gpu for GUI rendering. If you are using less cards that Your mobo could handle You could switch in BIOS Primary Graphic Device from PEG to IDE(or sth like this).