Bitcoin Forum

Hi guys.

I have a Rig with 4 XFX Radeon RX 580 8GB graphics cards and 2 XFX Radeon RX 580 4GB cards + Mobo Z270-A PRO + Intel Celeron G3930 + 8GB RAM + HD 120GB + EVGA 1000 Source G3 80 PLUS gold

I am currently mining dual ETH + Decred in Nicehash.

I guess the problem I'm having with the graphics cards is with the overclock.

The 4GB cards factory comes with:
CPU clock = 1366 MHz
Memory clock = 1750 MHz

The 8GB cards factory comes with:
CPU clock = 1405 MHz
Memory clock = 2000 MHz

With MSI Afterburner I setup 4GB graphic cards to the speed that it factory comes in CPU Clock and Memory Clock, and I did the same with 8GB graphics cards.

If the graphic cards are setup like I did they should work fine right?

I setup each graphic card with the speed of the CPU Clock and Memory Clock that each graphics card factory comes but the problem persists.

What profile or how to overclock recommend me to setup graphics cards?

What speed of CPU Clock and Memory Clock should I use for the 4GB and 8GB cards respectively?

Several times during the day I have to be turning off the system because the operating system get stuck and I must cut off the power source to turn on the system again.

I hope you can help me or collaborate me with your experiences.
Thank you very much in advance.
Greetings!!!

1) Get a better power supply, 1000W fore 6 of those GPU's without undervolting is inviting trouble
2) Use a tool like MSI Afterburner to undervolt the GPU's and then overclock the memory while keeping core low

Thanks for your answers.
Im going to undervolting with MSI Afterburner
Which values should I try in CPU clock and Memory clock in each graphic card?
Do you have any suggestions?

These clocks are way too high for mining.

Generally core clock of 1150-1200 is good enough, and alot of people would argue to set it even lower to reduce power and heat.
The memory clock depends on what type of memory you got; Samsung and Hynix can go for 2150-2250, elpida at around 2000 (never had a micron card, so I dont know about them).

Both core and memory would "most probably" work on 900mv; if your cards are good, they can work with even lower than that. that will give you around 100-110W reading from gpu-z in dual mining, and around 140-150w from the wall.

I personally use OverdriveNTool for overclocking and undervolting, and it works really well.

Check this guide, its really nice and hope its helpful.
https://mining.help/

99.9% the problem is that 1000W psu is insufficient for 6x amd dualmining unless you undervolt and underclock the gpu.  Even then, 1000W is barely enough.

My core clocks are around 1100.  Voltage is between 800 and 850.  Memory clock varies a lot you just have to test it, but +100 overclock on mem pretty much works for even the worst cards.

I'm running my rx580 nitro+ with core clock 1150 mhz and memory clock 2030-2050 mhz. I did a bios mod before and get about 30.5 mhs in claymore (mining ethash coins).

I have RX 570 (not 580) but the info below can give you some idea where to start.
RX 570 with Hunyx runs at 1150/1980 MHz (GPU/Mem)
RX 570 with Elpida runs at 1150/1950 MHz (GPU/Mem) - I had to lower Memory clock because graphics card was unstable on 1980 MHz.

Thank you so much for your answer.
I just setp up all graphic cards as it is explained in the guide you send me:
http://i0.wp.com/mining.help/wp-content/uploads/2018/01/ethereum_mining_windows_gpu_overclock_undervolt_template.png?w=906&ssl=1


All my graphic cards (4 from 8GB and 2 from 4GB) are set as it is showed in the picture above.
GPU Clock = 1150 MHz
GPU Memory Clock = 1800 MHz
Now Im testing the new configuration.
My graphic cards are RX 580 with Hynix chipset

your system is getting stuck because core and memory clock are very high or you undervolted a lot. Settings need to be balanced out. Also if you are not using igpu as boot then you need to make sure gpu0 is very stable.

Im running my hynix cards at 1150/2150, 880mv, and im getting 30.2mh/900mh eth/dcr
Those are RX 570,, 580 can do a little better with dcr.

Should I set in Memory:

  ... MHz      mV
P2   2150     900

Will It be ok?

and the rest as the same right?

Even for my two 4GB graphic cards?

you need to check memory errors and which gpus are crashing, memory errors can be deceptive too, also check incorrect shares, if you have any then you need to decrease the clock. Nothing will be gained by increasing memory clock if you don't do many tests if is okay to increase. Silicon to silicon are different and so memory chips, so if one gets to such overclocking that does not mean you will get to without issues. You need to read a lot.

Also many people complain about pool low hashrate and that is linked to stale and incorrect shares, stale linked to internet connection, incorrect shares linked to high memory clocks.

Thank you for your answer.
I dual mining ETH + Decred
Im using this pool for ETH: stratum+tcp://daggerhashimoto.eu.nicehash.com:3353
and this one for Decred: stratum+tcp://decred.eu.nicehash.com:3354

I get my payments in BTC for this dual mining

is it better getting payments in BTC or in each coin?

are these nicehash pools good for dual mining?

It is ok what Im mining?

Which pool would you sugest me for mining?

Thank you in advance for your answers.

I would recommend ethermine to mine eth, no point dual mining anymore. I would mine only ETH cause dcr is not even paying the electricity you use to mine it. Also dual mining stress your cards a lot and increases incorrect shares too cause it makes the pcb and memory chips of your gpu hotter. So turn dual mining off.

if your rig stuck a few times a day, usually it is not video cards issue or overclock. It is the riser. Disable video card one by one to find out which card cause the issue and change the riser. I have the same issue before and after i changed to a new riser problem solved.