6th graphics card not working in B250 Mining Expert motherboard

The manual doesn’t say anything about having to mix Nvidia with AMD unless you are using 14-19 graphics cards. Even then it says you can have 11 AMD and 8 Nvidia cards. Even so, if you can provide more specific tips I might try them out. Can you point me to any of those diagrams that show which slots are best for which cards?

1 Like

Theres a bunch of threads here on the asus b250 mining board, yes some with diagrams. You can review them but if you’ve tried everything else (sounds like you have) its a general consensus its your power delivery (PSU, house circuits etc.) or your board is defective.

I’m not the expert on 5700 XT’s but it frequently discussed here in the forums, with your issue and others. Its not an ‘easy’ card to ‘scale up.’ Also if you’re not doing more than 7-8 cards theres plenty of others out there that will suit your configuration better.

I have now tried all of the following without any success:

  • Updated the OS (including drivers and miners)
  • Switched to the beta version of the OS and updated it
  • Switched to Phoenix Miner (from TeamRedMiner)
  • Hooked up the second power supply more properly (using the equipment that came with the power supply kit for these two power supplies for this motherboard such that the two power supplies are hooked up to each other)
  • Switched two of the cards to be powered by the second power supply instead of just one. One of the cards from the second power supply works and the other doesn’t (still 5/6 working and one chosen seemingly at random to not work).

Try this:
All PCIe settings in the BIOS for Gen2 !!!
Connect the 6 cards in slots A01, A02, A03, A1(x16), A04 and B07; nothing different from that.
PSU at input A; make a new stick with the latest stable version of hiveos.
Start the system maintenance mode. Configure overclock settings. Disable maintenance mode. Restart system;
I have a single rig with the B250.
Works great with 12 GPUs (10x RX588 + 2x RX5700XT), 2 PSUS 1200w platinum in A and B;
The order of the slots must be this: A01, A02, A03, A1(x16), A04, B07, B09, B10, C13, C14, C15, C16;

02 days ago I added 2 RX5700XT, since then my headache started. With a fresh USB stick with latest stable (AMD 20.40) it recognizes 14GPUs and works perfectly, but if I turn off or restart the rig, it only recognizes 12 GPUs and i need insert a fresh system usb stcik;
I’m still trying to understand why. :frowning:

The additional slots for 14 GPUs I am using are A05 and A06.

I got it working! I had to enable above 4g encoding in the BIOS settings.

For the record here are the rest of my details of my current setup, not that all these details are necessarily relevant to getting it to work.

All the cards are currently in the C slots except the one with the HDMI out, which is in the 16x slot. I only did this because I was experimenting with it at one point because people with a mix of Nvidias and AMDs were recommending putting the AMDs at the end.

I switched the PCI-E to run at Gen 2 speeds instead of auto (BIOS setting). It may or may not have been necessary to make this change (probably not), but I was trying it as suggested by @amon49.

I am back onto the stable version of the OS, which has the latest version of OpenCL (AMD drivers). For some reason the beta version has an older version of these. I did not try this setup on the beta version, by the time I got it all working I was already on stable.

Thanks to everyone that tried to help and advise me!

1 Like

Good news!!!

Make sure you have the latest BIOS (1208).
You tried restart the system? Come back fast, without slowness and start the miner without high CPU usage?
This MB has a problem that when restarting the system it does not properly recognize all the cards or it has slow loading and high cpu usage.
The order of slots I suggested solves this problem.
As for the rest of the BIOs configurations, I confess that I’m not well remembered, but all PCIe Gen 2 and disable audio, sata, etc. Anything not to be used. And enable return in the on state whenever power goes out.
Today without electricity here for about 10 minutes and my rig restarted correctly recognizing the 14 GPUs. Hallelujah!! :slight_smile:
Good Luck!!

4g Encoding must always be enabled for GPU mining; I don’t have that motherboard but I think it has a “mining” option in the bios that enable 4g encoding and put all the pci-e in the right generation for mining; I think I’ve read that in the ASUS web page.

Naw everything checked out…

EXCEPT -
I tried unplugging the suspect and plugging into another, voila, problem repeated.
Something must be wrong on PSU, VGA rail buggered somehow., it ‘works’ tests fine and everything.
Down to 11 GPU’s, only one populated in A, B-Full, C-remaining.

I also ended up swapping the MOBO out and to my surprise it was not the cause. Whew.

I think I might be having that problem with the high CPU usage, but I’m not sure. I thought I could mitigate it by setting my HiveOS account settings to use a power cycle instead of a reboot. But now it seems to be rebooting every 30 minutes or so. But there are no obvious errors or anything that show up in the dashboard, HiveOS doesn’t make it obvious it is rebooting like that, so I don’t know how long this has been going on for.

Before changing the cards to different slots, I want to confirm it is getting high CPU usage. Also, switching to specific slots might help me for now, but my ultimate goal is to have 16 AMD GPUs on one board if possible. I have more coming in soon. Can you elaborate on what slots prevent this issue after adding more GPUs?

Shutdown, wait 10 seconds and start can help. I am also using this for 14 GPU.

As for restarting every 30 minutes, I start by reviewing the overclocking settings. Activate the watchdog and see the miner log. Will tell you which GPU dead.

Use six of the twelve slots I mentioned. I try the top 6 on the list or put two of the slots listed in each bank (A, B, C).

As for using 16GPU, I think it will be very difficult, I’m already having problems with 14. :frowning:
I don’t know a combination of slots for more than 12 GPUs. I am using additionaly A05 and A06 now.

If I found GPUs with prices close to the MSRP I would buy 4 more to make another rig with 6 and return ASUS for only 12 GPUs which works very well, but unfortunately I think it will not be possible in the near future. They doubled the price of GPUs. For now I will keep these 14.

PS: About 4g that you mentioned before, it is already activated by default in mining mode.

Thank you. Looking over your previous message I am now seeing that you did also mention a 12 slot combination in addition to the 6 slots, so I appreciate you bringing this to my attention. I will try it out to see if it alleviates my issues.

After looking at the output from top, I don’t seem to have any issues with high CPU usage. Either I never did or the setting to do a full power cycle instead of a reboot is preventing it.

I am still having the reboot issue, and apparently it is happening about once per hour instead of once every 30 minutes as I thought before. It is strange because there is no miner restart alert on the dashboard. Normally if that happened I would click it and I could see the log leading up to the reboot including which GPU died. I also tried poking around in the file logs on the file system, but the logs only show me what happened after the reboot. This is the case for every single log file I looked at. It is as if the logs are wiped upon reboot…

As far as the mining mode goes, you are not the first person to mention it to me, but I turned it off a while back. It prevents me from having access to all the settings and there was an important setting I had to change with mining mode off. I think it was the one to power on the rig automatically upon power restore, which is a really important setting for me. I’m not sure if I turned off 4G encoding at some point a while back or if it turns off when I disable mining mode. None of this matters really, just an FYI.

hi
check your conection
problem in your connection
are connect molex to pci-e power ?
are update mainboard bios ?
try to use onboard graphic and send your first boot screen
annnnnnnnd if you use 2 power 1 power for start main board and one bitcoin power to graphic 8 pin you most start first 8 pin powers then start mainboard power

Thanks for the responses, everyone. It was really helpful in getting my B250/rx5700 rig going. I made it past the 6th card threshold. To those still seeking possible solutions: In the end, I just needed to increase virtual memory. My virtual memory was only big enough for 5 cards, so only 5 would run.

how do you change the pcie gen setting in the b250 bios?

This topic was automatically closed 416 days after the last reply. New replies are no longer allowed.