Asrock h110 pro btc cant disable CSM

[ooskie]

I’ve had one of these ASRock H110 Pro BTC+ running with 12 GTX1070’s for over a year.
Had a few of these Bios settings different, but since I re-pasted the heatsinks on 5 of them it will only run 11 cards.
Tried changing all settings to match these except CSM (will not boot with CSM disabled) with no change in symptom: It boots ok and recognizes all 12 cards, but once it starts mining with the 6th GPU, The 6th GPU fan goes to 100% and the rig only runs with 5 GPU’s. I have the same power supplies and Bios. haven’t changed anything except heat paste. Have tried swapping several risers and all work fine with other GPU’s and slots, but only with 11 cards.

It does not matter which GPU I remove or use in #6 slot, and #6 slot works fine when I have 11 cards.
Been a week or two like this and I’m ready to try downgrading to an early April build, or maybe this MB has something weak.

ASRock H110 Pro BTC+
Intel® Pentium® CPU G4400 @ 3.30GHz
4GB Ram
M.2 SSD ATA WDC WDS120G1G0B- 120GB
1000W evga supernova p2 and 850W Corsair RM850x

Might not be related, but LA seems to fluctuate a bit from 0.6 to 2.4
Any one have any thoughts?

[Smining570]

thank you for instructions. I have noticed that my monitor wont work if 1 gpu is connected to x16, “Connect just one GPU to the x16 slot (longest one).” , so is it mandatory to have one gpu connected to that x16 slot and if so, do these bios settings change the monitor output from x16 so that there will be hive running on the screen from onboard graphics ? …or is the rig managed only remotely after there is +9 gpu´s connected? I do understand that it was stated that rig will only boot with monitor connected, but will there be output when x16 have gpu connected? …sorry, im a bit confused about this so just want to understand what i need to aquire to run 10gpu´s on this board. thanks

[Smining570]

plastic playing cards are really thin and strong, wery good for insulation between the ports… cutted to a bit smaller pieces naturally

[Cinco29]

A tiny piece of electrical tape works wonders for me.

Have 8 of these MBs all running 12 gpus and have yet to have a single problem. I did originally have issues trying to get 13 cards to mine but tried it with leaving the 16X slot empty and it worked. It has worked on all the other boards the same since then.

Give it a try and hopefully you guys have some luck!!

[Smining570]

have u 12 gpus with default bios settings,? i have 11 gpu now and i Dont see why the board dont want to find card 12…all connected to small slots. weird board…

[Kpereira]

Ive tried these settings and they work up to the point where all the 13 GPUs are recognised, however get a lot of errors once the miner starts.

Msg issuing pre-check failed and smu may be not in the right state

amdgpu failed to read reg ec48

And few others as i keep trying to troubleshoot.

Have changed risers, have changed OC settings, miners, tried beta image, back to stable image etc. Nothing seems to work.

Im using 13 AMD cards (3x6800XT, 1x5700xt, and the rest are 6700xt)

Any help would be appreciated.

[Baduma]

Holy molly. You, sir, are my hero.

[ihg70]

Hello again after a long time

I’m glad that instructions I prepared helpd most of you. I wish I could help more for recent posts but I’n not using Asrock pro H110+ anymore. 2 months ago it’s burned One of PSU cable had faulty and it burnt both motherboard and CPU. Other components were OK (SSD, RAM).

So I had to buy new mobo.

Asrock was my first mobo to start mining. I had chosen it becaus to reduce initial cost like buying 1 CPU, 1 SSD, 1 RAM etc instead of 2.

But later I saw it’s not a good idea. 6x or 8xPCIE motherboards are far better than 12x, 16x or 20x PCIE motherboards.

• Most important aspect is less operation down-time. If 13x mobo went reboot, you lose 13 GPU’s operation. It’s really big loss when you try to tune your GPUs. Every reboot means you restart all 13 GPU and no mining at all during this restarts. But with a small motherborad you lose only 6 GPU time.

• 6xPCIE mobos has simpler operation. Less cables allow easy intervention if you need to change something, for replacements or swaps. And easier troubleshooting.

So when I set my second rig I went for Biostar with 6xPCIE slots. And for other rigs I added were either Biostar or Gigabye motherboards. All are working good so far with less down-time and with very easy operation &maintenance & interventions.

After burnt AsRock, I bought a new Biostar mobo with 8xslot. It’s good as well.

I have 7 rigs with 39 GPUS since January 2021, (see the picture below).
The last Biostar mobo with 8xslots, I only could fill 3 slots and 5 more slots to go to make full rig.
All my GPUS are RX5700 XT with different brands, mostly Sapphire and Gigabyte 5700 XTs.
Also I have 3 Asus 5700.

My total minging hashrate is 2.06 GH/s with power consumption of 3400 W
Avarage hashrate per GPU= 53 MH/s
Avarage watt per GPU= 87 watt

My total investment is (includes everyting) ≈ $25 K
From Jan to end of May I mined around 13 ETH and got $42 K (avg. 8400 / month)
In this period, my total electric bills were $1700. Monthly average around $350

[silence23]

Hi, here is beta BIOS from ASRock support with disabled CSM in default settings. I had same problem with more than 9GPUs, now everything works well.

ASRock H110 Pro BTC+ beta BIOS L1.64 with disabled CSM by default

[Smining570]

hi. Your power consumption for +2GH is amazing low, is that measured straight from the wall, i had a while ago 2 rigs with 850mh and consumption was over 3KW, i wonder what am i doing wrong here…the difference is huge…

[Smining570]

Here´s is the Beta Bios all who want download it.

BETA BIOS

[Dantheminer]

Wow !!! man thank you so much for coming back with this. flashing bios right now!

[Dantheminer]

@silence23

BIOS from ASRock support with disabled CSM in default settings. I had same proble

My Goodness Gracious! This is Christmas in June!!! If you happen to be in california, I will buy u a beer!

[ihg70]

@Smining570
It should be at least 50% more on wall. I did not measure.
I lowered it even less than actual state but I could not keep them stable. Now all my cards and rigs are quite stable for a long time after a long tuning effort at the beginnings.

As you can see from the picture below, they work days and days without any restart.

[jolafriite]

Hi ihg70

May be you have the solution for my problem.

I just added a new card to my rig and since I have some cards who drop in hashrate (and also in W) a few minutes after the reboot.

PSU1, 1050W, GPU 0, 1 & 2 => 3080 (98MH), risers, Mobo, CPU, … for a total of 770W.

PSU2, 1200W, GPU 3 => the new added 3080 (98MH), GPU 4 & 5 => 3070 (63MH), risers for 480W.

With the new 3080 in PSU2 (GPU3), GPU0 and 2 of the GPU1 drop to 89 (and sometimes lower) and 90 for the second instead of 98MH/S !!! After they stay stable at those hashrate.

Where can the problem come from? (no problem of T° because they do not change before/after)? How to solve this?

Mobo H110 pro BTC+; extra power connectors connected with independant SATA, each risers with its own SATA, as well as for GPUs.

[marctheshark]

So I can confirm you dont need the molex’s connected and you don’t need to reset the BIOS. Though I am able to get all 13 running, it will reboot and occasionally the csm will reset, don’t understand as to why

[Grea]

Did you try the beta BIOS linked in this thread with CSM disabled by default?

[oskuro]

I did this tutorial before: Asus B250 / Asrock H110 e1000e NIC hangs entire network [SOLUTION] but I’m still losing internet connection.

I have applied your Bios settings, let´s see if that solves my issue.

Thanks.

[oskuro]

Still happening. I don’t know if it´s the damned asrock h110 or what. I will switch to windows or maybe I just remove this mobo and sell it. I’m tired of this shit.

[Grea]

I only have 7 GPUs on mine and it runs for months at a time. I have GPUs to fill it up, but dread creating problems where there are none today. It was a pain to get it to 7, remote upgradable, and stable., so I am likely going to add GPUs to another system myself.