ASrock H110 Pro BTC+ 13 GPU Board

[Vann]

1% max is normal. You can lower the rejected share rate by lowering the memory overclock, but the hash rate will also go down. Network latency is also a factor, so you should mine at a pool with a server closest to you.

[1714850786]

1714850786

[1714850786]

1714850786

[1714850786]

1714850786

[1714850786]

1714850786

[ss890]

Looks great to me. Being a gpu miner that would be dream motherboard but I'm not sure how they gonna set it up with the core. The motherboard must be cooled properly otherwise it will get burnt in an hour or so. That's really not funny as it can happen with normal 6 PCI card support too.


If they have planned to make something like that then they might have thought of it as well. They should consider the overload capacity of ram too because more the cards the more memory it will need for single motherboard. Wondering how the bus lines on the mother board will react with so much of the hash at single time. It better be perfect on all the dimensions.

[ImXotep]

I use Ethos 1.2.5 and optiminer 1.7 with 12 RX580 on ASrock H110 Pro BTC+ . Every 9 hours like by timer powertune set from 3 to 7 and cclock from 1250 to 1366. Where problem and how I can fix it?

[jang430]

Anyone using this board, with Claymore, the the configuration as I have?  5 GTX 1070s, and 5 RX 580s on the same rig? 

Whenever I use 5 rX 580s and 5 GTX 1070s, Claymore will hang as soon as it starts but 5 RX 580s, and 4 GTX 1070s, it goes on without a problem.  I then change my virtual ram from 16 GB to 32 GB, then proceeded with addition of 1 more RX 580s, for a total of 6 RX 580s, with 4 GTX 1070s, it will start mining, then after a while (an hour or less), the machine needs a reboot since it's no longer responsive, and I can see nothing in the screen.  Remote access to this machine doesn't work either. 

Using Windows 10, 64-bit, Claymore miner 10, AMD Blockchain drivers Aug. 23, and latest Nvidia drivers (forgot the version).  The machine has 2 x 4 GB RAM.

Any ideas?  Can't get all 11 cards onto this board and run without problems inside Windows 10.

[Wefud]

what does everyone consider a normal acceptable level of share rejection? 1%? I have 13x rx580 on this board

What operating system do you use?

[Ursul0]

Looks great to me. Being a gpu miner that would be dream motherboard but I'm not sure how they gonna set it up with the core. The motherboard must be cooled properly otherwise it will get burnt in an hour or so. That's really not funny as it can happen with normal 6 PCI card support too.


If they have planned to make something like that then they might have thought of it as well. They should consider the overload capacity of ram too because more the cards the more memory it will need for single motherboard. Wondering how the bus lines on the mother board will react with so much of the hash at single time. It better be perfect on all the dimensions.

This is very confusing. Particularly these parts: "how they gonna set it up with the core" ,  "motherboard must be cooled properly", "overload capacity of ram", "bus lines on the mother board will react with so much of the hash" WTF man?

EDIT: I've got some of those arriving in a few days. Will see how it goes.

[davemanet]

I have 7 gpus working fine, but when I add the 8th one...... my other GPU's start declining.  I'm using 4gb of memory.  Anyone know what might be happening?

[jang430]

@davemanet, you're limited to 8 of a kind in Windows. 

As for me, I don't know why my 5 GTX 1070s doesn't allow me to accept more than 3 RX 580s.  Don't know how to fix this.

[jang430]

I have a bizarre issue that I've been fighting all weekend. Have this board with a celeron and 8gb of ram with a m2 ssd, runs fine with 6 1070s attached to a 1000w psu that also powers the board (900w at the wall) then I tried to add 2 AMD cards... I used a 750w psu that was powering those same 2 (RX480 w/ BIOS mod) cards with a Y cable for turning it on and everything booted up and ran fine for about an hour, then one of the AMD cards dropped out (Claymore dual miner) and the system restarted and that card no longer showed up in device manager. I swapped out the riser and no change, connected to different mobo slot no change, tried many many different things and the one card never lights up or fans spin up like its not getting power. I swapped the 480s out for 470s and it does the same thing, I put the cards back in the machine they came from with the 750w psu and all works fine, I've never run a rig with 2 psus before, is there something I'm missing? I'm powering each riser and the card connected to it from the same psu

@bucketofsocks, I think it's somewhat similar to my case, see below.

Whenever I use 5 rX 580s and 5 GTX 1070s, Claymore will hang as soon as it starts but 5 RX 580s, and 4 GTX 1070s, it goes on without a problem.  I then change my virtual ram from 16 GB to 32 GB, then proceeded with addition of 1 more RX 580s, for a total of 6 RX 580s, with 4 GTX 1070s, it will start mining, then after a while (an hour or less), the machine needs a reboot since it's no longer responsive, and I can see nothing in the screen.  Remote access to this machine doesn't work either. 

But sadly, no one shared insights on this yet.

Jason

[jang430]

You would think after all this mixed card BS & hassle with Windows people would just use Linux. I don't get it.. To those using windows: If you just tried Linux you'd never go back

Have been wanting to go Linux after all these hassles.  But I've already got 5 GTX 1070s, and 6 RX 580s.  So which OS support mix mining?  ETHOS says it doesn't support it.  Any update on this?

Thanks

[Vann]

I have a bizarre issue that I've been fighting all weekend. Have this board with a celeron and 8gb of ram with a m2 ssd, runs fine with 6 1070s attached to a 1000w psu that also powers the board (900w at the wall) then I tried to add 2 AMD cards... I used a 750w psu that was powering those same 2 (RX480 w/ BIOS mod) cards with a Y cable for turning it on and everything booted up and ran fine for about an hour, then one of the AMD cards dropped out (Claymore dual miner) and the system restarted and that card no longer showed up in device manager. I swapped out the riser and no change, connected to different mobo slot no change, tried many many different things and the one card never lights up or fans spin up like its not getting power. I swapped the 480s out for 470s and it does the same thing, I put the cards back in the machine they came from with the 750w psu and all works fine, I've never run a rig with 2 psus before, is there something I'm missing? I'm powering each riser and the card connected to it from the same psu

@bucketofsocks, I think it's somewhat similar to my case, see below.

Whenever I use 5 rX 580s and 5 GTX 1070s, Claymore will hang as soon as it starts but 5 RX 580s, and 4 GTX 1070s, it goes on without a problem.  I then change my virtual ram from 16 GB to 32 GB, then proceeded with addition of 1 more RX 580s, for a total of 6 RX 580s, with 4 GTX 1070s, it will start mining, then after a while (an hour or less), the machine needs a reboot since it's no longer responsive, and I can see nothing in the screen.  Remote access to this machine doesn't work either.  

But sadly, no one shared insights on this yet.

Jason

I was having similar problems when adding two Nvidia cards together with 8 AMD RX series cards. The mining would crash after about 1-2 hours. I think it's because the H110 chipset only supports 6 native PCI-E lanes. The rest of the PCI-E slots use an expander chip, so it's important to use the same type of cards, either AMD or Nvidia, in each of the PCI-E slot pairs to prevent conflicts.

Ultimatley what solved it for me was moving the two Nvidia cards to the last PCI-E slots on the board, 6_1 and 7_1. In the Bios I also set the all the PCI-E slot speeds to GEN 1, except for the Nvidia cards to GEN 2 (they wouldn't be recognized if set to GEN 1) and I also set the x16 slot to GEN 3. I then removed the drivers in safe mode using DDU and reninstalled the drivers. I installed the AMD blockchain drivers first and the Nvidia 385.12 drivers last. Runs stable now since I did that.

[jang430]

Thanks Vann, I hope to try this out asap, and get favorable results.  As it is, I have 11 cards, 5 GTX 1070s, and 6 RX 580s, and not being able to get them to work is not only highly frustrating, it's also costly. 

My other question is, seeing all this talk about simplemining, and ETHOS, does it require card to be grouped such as posted above?  Or is this issue with Windows 10 only?

[Vann]

Thanks Vann, I hope to try this out asap, and get favorable results.  As it is, I have 11 cards, 5 GTX 1070s, and 6 RX 580s, and not being able to get them to work is not only highly frustrating, it's also costly. 

My other question is, seeing all this talk about simplemining, and ETHOS, does it require card to be grouped such as posted above?  Or is this issue with Windows 10 only?

Mixing AMD and Nvidia type cards is only supported on Windows. On Linux you can use up to 13 AMD or Nvidia cards with this board, but they must be the same type for all the cards.

[jang430]

SimpleminingOS seems to show they allow both on same rig.  Though I cannot confirm as well.  Will try your suggestions first.

[Metroid]

Mixing AMD and Nvidia type cards is only supported on Windows. On Linux you can use up to 13 AMD or Nvidia cards with this board, but they must be the same type for all the cards.

So for the asus b250 mining expert would be the same, 19 amd or 19 nvidia cards, as long as all matches the type. The problem is if one card breaks and cant be replaced, you have to use 18 cards and so on.

[jang430]

@Vann, are you currently able to change CC, mclock, power limit with your AMD and Afterburner 4.3.0?  I'm using AMD Blockchain drivers, but I cannot.  I can only change settings for GTX 1070s.  Do you know if EVGA tool can do this?

[jang430]

BTW, does it mean this motherboard doesn't require 2 molex inputs if you disable it in the bios?  I had to use a separate PSU just for the motherboard since I'm using server psu with pico board.  Kindly confirm this.

[BenScoobert]

what does everyone consider a normal acceptable level of share rejection? 1%? I have 13x rx580 on this board

What operating system do you use?

SMOS

[msmde]

I was having similar problems when adding two Nvidia cards together with 8 AMD RX series cards. The mining would crash after about 1-2 hours. I think it's because the H110 chipset only supports 6 native PCI-E lanes. The rest of the PCI-E slots use an expander chip, so it's important to use the same type of cards, either AMD or Nvidia, in each of the PCI-E slot pairs to prevent conflicts.

Ultimatley what solved it for me was moving the two Nvidia cards to the last PCI-E slots on the board, 6_1 and 7_1. In the Bios I also set the all the PCI-E slot speeds to GEN 1, except for the Nvidia cards to GEN 2 (they wouldn't be recognized if set to GEN 1) and I also set the x16 slot to GEN 3. I then removed the drivers in safe mode using DDU and reninstalled the drivers. I installed the AMD blockchain drivers first and the Nvidia 385.12 drivers last. Runs stable now since I did that.

[im]htps://image.ibb.co/eTxFKQ/Screenshot_from_2017_09_23.png[/img]

Do you have any thoughts on using a tb250 btc pro over a h110? i have been reading that people are having some issues with the h110. nothing major, just a difficult time getting it up and running

[jang430]

You know what? For the first time I got all 10 cards on a same rig running. Special thanks to @Vann for pointing out the pcie groupings. Following your advice solved the problem instantly.