some facts.
you can expect around 30,5mh/s x 6 gpus from every rig (that makes around 183mh/s). You can expect something around 135W for every card. so around 22kW of power considering all the gpus + components.
Your ETH hashing power should be around 4800mh/s. If your Electricity cost is around 0,10 you make 511$ right now (that's not even considering ROI, bill taxes, or government taxes over profit. 0,11cnt/kW you make 352$ right now. 0.12cnt/kW you make 194$. and at 0,13cnt/kW you make 36$.
Now, the average price in my country for every kW is around 0,15€/kW (we pay 26% taxes over electricity prices due to leftists "idea" with green energy). Which is around 0,17$ (including all taxes). Talking in numbers, a guy living in my country with your setup, will be losing almost 600$ per month at current prices.
I can add some information, that can help you to collect some more facts.
The average price in my country - less about $0.06/kW.
In the first post TS shows Mh and power consumption - totally it 6454/21910. And $0.1 per kW.
Less profit - much miners turns rigs off - difficulty down - more profit.
More ETH at low cost and low difficulty - HODL - more profit.
Give some more facts please.
you seem to be like one of those poor guys outthere that would kill for 10$. Sorry, i don't want to deal with people like that. It makes no sense. Ohh, BTW, at 0,06$/kWh you live in a third world country, so i undestand your way of thinking. You make more mining than working, so 10$ for you can be way more money than working. It makes sense. For 10$ profit per month, i don't even take the cotton sheets from the equipment.
Oh, BTW, at lowest prices on ETH (80$ more or less), it never went down of 134.000Gh/s. Even if we have seen a drop in value of over 35% in last month, the hashrate went from about 145.000gh/s to actual 172.000gh/s. In last two days, with masive drops of nearly 25%, it increased from 162.000gh/s to 176.000gh/s even with bloodbath in the streets.
So people turned on their equipments even at heavy loses, which show how "intelligent" are miners nowadays.
That's because the difficulty is a lagging indicator. When the price first started jumping to $250-300 difficulty actually DROPPED. However, as the price increased and stabilized for a couple of weeks difficulty increased. I suspect if this price drop continues or stabilizes around $160-200 we'll see about 2.0T difficulty, but it takes time for the farms to react.
Further updates
I've been busy trying to stabilize my rigs, 90% of my problems involve riser issues and cards "falling off the bus" the rest is just the typical overclock/undervolt issues and finding the right settings. I have to praise SRB BIOs mod. It literally saved me as 'one-click' PBE was awful for these Hynix cards. Resulting in tons of crashing and BSODs. Once I tried SRB it increased my hashrate for Hynix and Elpida to 30mh/s per card at 1950-2000mhz@860-875mv(core/mem) which is about 120w~ a card. Total hashrate for my rigs is around 385-393mh/s for 13 GPUs @ 1650w~ total power per rig.
I've been mining since 2017 and one thing I learned is to HODL the coins you earn as much as possible and never ever sell during down-periods. Right now while my rigs "aren't as profitable" due to the down-turn. In one month time that can literally change as markets pick back up. I cover my electric as much as possible with my normal IT job and sell coins to cover the rest.
As of the time of this writing, since I started this thread and my farm back-up mining ethereum in Ernst I have collectively mined 9.5 Ethereum. Mind you I was slowly getting rigs online and lots of crashes so this hasn't been 24/7 mining. This is one-man operation as well and I have a normal 9-5 job, girlfriend, and life. I spend as much time as possible working on my farm as a second job. My estimation is to achieve 90% stability by End of July. Right now I'm at 65% with 11/14 rigs running continually but with a few crashes/restarts due to GPU crashes in claymore.
mmm....big farms react almost instantly. changing their algo can be done almost automatically with their software. There are people outthere that can switch coins with one click. Moreover, they can also tell the computer to switch to certain algos during certain times. They load profiles directly on the miner (not bios) and as soon as they load the miner, the program gets the card working.
I myself was able to switch from ETH to ETC miles away from my house when i was mining back in 2017.
Now..... i don't normally help people with issues, but i´ll give a few tips in case they can help.
the sweet spot for RX 470-570 (mines were 570's ) is: 1150 mhz on the core, 2100 mem. AFAIK 470 and 570 are almost the same card, but 570 is a bit better, to be honest.
My experience is that you should not undervolt the memory. My Sapphire RX570 nitro+ cards did crash a lot if i undervolted the memory, so i left them at 1000mv stock. 850mv on the memory...well...maybe it's ok for 2000mhz, but take into account that RX 570's nitro+ at 1750 stock mem clock had 1000mv stock.
Now, the voltage.
i have 3 RX 570's nitro+ 8gb
1º 1150/2250 831-837mv core/ 1000mv mem 31.4mh/s
2º 1150/2190 831-837mv core/ 1000mv mem 30,8mh/s
3º 1150/2125 831-837mv core/ 1000mv mem 30,3mh/s
They were stable for months at this settings. I tested for two days all of my cards before going into the mining rig. i programmed the bios in all of them and i tweaked as much as possible the voltages except for memory. ETH is a pretty intensive algo with the memory, and i prefered to forget about having problems with undervolting. I suppose your undervolt saves you some money.
my 2x6 mining rig are off. and they will continue to be off for a large period of time. I will only turn them on if i can get 100$ profit per month or more after electricity bill, which i think will not happen this year. At current prices , and the very expensive cost per kW i would be losing around 15 $ for month with bills of around 350€ or a bit more in summer due to fans having to work hard in my room to deal with heat, specially where i live.
So....yeah...mining is not worth for me.
The person who sold all these 150 gpus was relieved from heavy sleepiness duty ehhe
He sold and took a long vacation and never looked back. This person was finally unchained.
You trolls will never understand what freedom really is until you do exactly what this person has done ehhe
This person enjoy and understands what freedom really is ehhe
The funny thing is that this is how I felt when I ran a large GPU farm. I still mine but with a few GPUs here and there and mostly for fun. Basically only betting on ETHBTC going to 0.05BTC one day and will sell it all for BTC and Hodl that.
I remember the stress of having GPUs crash and I had alerts set and was awoken many hours of the night. And when I wasn't home I kept getting alerts about crashed rigs and had to drive home, reboot the rig, and drive back. This got very stressful during the $5/day/GPU when there was alot of profit on the line.
Now that the profitability is much lower you can sleep throughout the night and not worry, because a crashed rig doesn't cost you as much as before.
Luckily I don't have to drive anywhere since it's in my garage simply a walk. However I do still get a little stressed or annoyed because when they do go offline I don't know why, which is the big problem. I very much just want to achieve some stability but right now I haven't gone a full day without at-least one rig going completely offline :\.
something should be wrong in your config. both of my 2x 6 RX 580 were using 850W from the wall at 1150mhz 850mv more or less and 1000mv mem. with 13 gpus you should be using around 1850W rock solid, not 2000W as you said. That is due to PSU ineffiencies. It depends on the type of PSU you use, the power it provides, if you have both of them linked and so on.
Now, the point. You can undervolt your mem with this ideas.
1º you want less power consumption, not performance. So you go for the same voltage and same mem clock with all cards. That is at the cost of performance, so less mining power. My cards were tested individually. So, i would never apply the same numbers for my cards not considering i can extract a little more from each one. i know for a fact that all of them are giving the max performance at the lowest possible power consumption.
this are the numbers
401mh --- 1850W 59.10$ (profit) 55% 0.02682 (mined day) 0.8047 (mined monthly) <-------- my config
390mh --- 1700W 63.19$ (profit) 64% 0.02609 (mined day) 0.7826 (mined monthly) <-------- your config
This are average numbers.
This means that, in a investment to profit ratio, your undervolting makes sense. You use a bit less money on bills, but at the cost of mining less ETH per month than me. Who do you think is going to get more profit if the coin goes up in price?? And, on the contrary, you would win because you would be more power efficient which is very good when prices are low.
At a certain point, if the price scales, the more you mine, the more you make because that extra mined coin will be totally worth the extra electricity. In fact, i have two bios for my cards. One at 1150mhz and another one at 1240mhz, which adds like 0,3-0,5mh/s to each card.
with my config, with 13 cards, i would be doing like 406mh/s. With my 2x6 rig (12cards) i was doing something around 376mh/s. I would never be able to obtain those numbers applying the same config for all cards.
by that time, i prefered to have more coins rather than using less electricity because it was worth. Now...well...my config would be more expensive to maintain. But if coin goes up, it would be worth that extra power.
Anyway, you can just stick with 850mv and deal with crashes or increase the ram voltage and forget about having to get the car at night to solve the issue. That's the difference.
