у вас 12/13 красных карт на двух БП или на трех?
4 БП, от 650 до 800 Вт. Один БП на 2-3 карты с райзерами.
Всем привет. Установил 6×gigabyte 1660s, 5 работает 6ую плата не видит, ругается красным на нее, хотя по схеме в инструкции вроде такое именно подключерие должно быть. Установил последовательно в первые 6 слотов. Райзер к 6ой карте менял. Карту местами менял. Что может не так быть, что проверить? Спасибо. Блок питания на 1800w, серверный
Апдейт. В итоге 6ую карту увидела hive os, карта шуршит там по монете, те работает, но при загрузке материнка не отмечает 6ую карту зелененьким… такое вообще возможно?
драйвера обновите
I recently set up a rig consisting of Asus B250 Mining Expert and 6 MSI Armor Rx 570 8Gb GPU’s. I had huge problems getting the system going with the latest version of Hive OS. Updated the motherboards bios to latest version, but that did not work either.
I think I finally found out what the problem was. I was using the PCI Express x16 slot (the slot you can attach a GPU directly to). When I used a x1 slot instead the system started working flawlessly. I do not know why I cannot use the x16 slot. Is there a bug in Hive OS?
Set PCI-E Link to Gen2. This option exist this option exists separately for x16 slot and x1 slots
Подтверждаю. 12 карт 588 работают. Теперь риг умеет ребутаться не только по питанию. 2 года искал решение, ставил карты, как в инструкции от матери (в перемежку с нвидиа). Ничего не помогало. Давно забил и поставил на вайфай розетку. Сейчас, случайно наткнулся на эту тему. Просто переставил карты как на фото. Настройки в биосе так и стояли, как тут в описании. Все работает! Несколько раз ребутал удаленно. С первого раза - все онлайн. Без отключения по питанию! Не знаю, как вы до этого допетрили, но огромное спасибо! Лучше поздно чем никогда.
The motherfuckerboard will not work if your GPU overclocked high.
I did this on a rig based on the H81 Pro BTC R2.0 ASRock motherboard. And this seems to work. Rig has been running now for almost 20hrs. So Gen2 seems to do the trick. Thanks!
Hello ! Thanks for sharing this info.
Sorry if the question was already asked but does it work for rx 5700 xt ?
My 12 cards are detected on this mobo but rig is unstable. I put Gen 2 but when I did amd-info, it shows Gen 1 and Gen 2 on some PCI-E.
Should I put my cards like you show on this pic or it’s an issue with my mobo ?
Thanks for any answer.
This is a late response but some folks may find it helpful…
TLDR; on Win10 I can get 14 x Nvidia cards to run with no issues and only 9 x Nvidia on HiveOS
Up until recently, when switched to HiveOS, I’d been running to two identical Win10 B250 rigs with 10 x 1070ti NVidia cards with no issues. When I had to send one of my B250’s in for service, I successfully ran 14 1070ti Nvidia cards with no issues. But, if I added the 15th, the system would not load.
So, in short, on Win10, the max number of Nvidia cards is/was 14 cards.
Now onto the HivOS configuration
On HiveOS, no matter which slots (i know because I tried them all in various configurations) I use, the max number of Nvidia cards I can have on a B250 board is 9.
But to you point, ASUS officially only supports 9 Nvidia cards. I just found it interesting that with Win10 i could get 14 Nvidia cards running with no issue.
Hi guys,
is there any reference for the addressing of PCie ports?
I’m trying to keep log of individual cards. I expected that solid card reference should be by the PCIe address (format xx:00.0). however when i add new card, some of the cards keep the same adderess but some of the change them.
Thanks for your reply
ondrej
I have this on one of my rigs, granted I’m only using 8 slots because that is all the frame will accomdoate.
I had massive problems with this until I reflashed the bios back to the very first version. Everything then suddenly worked.
Hi guys , i have this card but have a lot of problem with .
i have a rig with 5 NVIDIA RTX GPU’s (4x rtx3070 + 1xrtx3080) .
when it run under 5 rtx gpu’s the rig is so stable but when i try to add 1 new amd card (rx6800) the rig don’t stop to crash and restared
can someone help to fix this
hi bro , can you send the link and the way to flash the card ?
make sure you enable 4G and mining mode.
I think it was v311 from that I downloaded to a USB, booted into BIOS then reflashed. https://www.asus.com/us/Motherboards-Components/Motherboards/All-series/B250-MINING-EXPERT/HelpDesk_BIOS/ I would consider myself an experienced novice but even I managed it. Basically I rolled back the BIOS as far as I could… I’m not interested in fancy support for latest intel processors etc
I have experienced the same results although I have been able to run 14 NVIDIA GTX 1070 GPUs both on Windows & HiveOS. Not sure how some people get 15 - 16 GPUs considering the amount of Intel CPU PCIe lanes. I have been experiencing intermittent disconnects with 1 GPU on Windows & HiveOS.
The difference between Windows & HiveOS is that Windows would continue mining. HiveOS will lose the remaining connections to each GPU, which could only be fixed by either scenarios:
- Switching to different mining software then revert back
- A reboot
I would normally see this error message while mining ETH with Pheonix Miner before any issues with the GPUs:
CUDA error in CudaProgram.cu:388 : out of memory
Although I’m using cards that have 8GB of memory. I’m not exactly sure what is causing this.
The watchdog isn’t helping. It’s been restarting the mining software nonstop but not the rig.
Attached screenshots to show what I’m explaining.
can you show me how you do for put card in the pcie x16 with the risers around it ?
thnaks
I have switched to this configuration and have been testing it for approximately 3 weeks now due to random GPU disconnects. I’ve figured out one card of mine does not do well at a certain power limit. This PCIe slot configuration helped me realize this.
You can insert a PCIe x1 into the beginning of the PCIe x16 slot. It’s compatible, just uses less lanes is all.