Crypto currency mining

[ainarssems]

In my experience, it's the other way around. Using NiceHash, both AMD and Nvidia cards will usually show about 8-10 compatible algorithms, however, it seems that most of the time NiceHash chooses to mine Ethereum on the AMD cards, occasionally switching to Zcash. With Nvidia cards, switching algorithms seems to occur much more frequently, with NiceHash selecting more algorithms/coins such as Lbry.

That's just something I found on web so could be wrong or could be out of date. Also apparently latest generation Pascal cards can do more than previous generations so best to stick with GTX10xx cards if you are going nVidia and not lower that GTX 1060 6GB version.

When I do benchmark with NiceHash there is only 1 algorithm for CPU in my HP Micro Server Gen 8, there was 4 or 5 on GTX 760 and 13 on RX480. Oh well I forgot to say that I received my first RX480 this morning and have been doing some testing. That thing is massive, I had to cut out HDD cage from the case to make it fit.

It turns out GTX760 was very power hungry at idle PC using 150-160W, with RX480 it only uses 90-100W but a bit more while mining. PC with GTX was 270W, RX is using 306-308W while mining. I might tweak clock speed and voltage settings to see if I can reduce power consumption. Very interesting considering that GTX had 2 power connectors 8 pin and 6 pin but RX only has 1x 8 pin connector beven if RX is using more power at full load. It could also be that be that CPU is using more power and not the card itself as with GTX it was only about 5% but now is about 15% CPU load. RX is also generating a lot more income, GTX was about 0.0002 BTC/day, RX is 0.0016 BTC/ day so x8 more while using 1.14x more electricity. I might also get away with my current 550W PSU and adding second card. PSU does gives 41A on 12V single rail but it might get loud at that load due to fan speed. I measure power consumption at the wall while it looks like that power stated on PSU specifications is output power 550W and online reviews show it consuming 640W at full load outputting 550W, 492 of them on 12v rail while providing stable voltage. I think with 2 cards at full load I will be about 500W from wall.

With single card I am looking at £65-80 per month at current BTC price while the PC is consuming about £24/month. With 2 cards it should be around £130-160 per month with £39 in electricity costs.

Another issue to address is how to wake the second PSU. You'll either need a relay driven by the first PSU to connect the appropriate terminals or you'll need an adaptor like this: http://www.ebay.co.uk/itm/Dual-PSU-a...-/191568455861

You just connect green wires on 24 pin plugs, it just gets connected to ground when you switch on PC. I suppose PSU wake up time could be issue when using 2 different PSU's so maybe best to stick 2 identical which should fire up in similar amount of time.

[moltuae]

I thought the RX 480 cards were big too .... until I saw a GTX 1080 Ti card!

The trick with the RX480 I found, is to lower the GPU speed by about 15-20%. At around that speed, the power consumption drops by about 50W with very little impact on performance.

Originally Posted by ainarssems

You just connect green wires on 24 pin plugs, it just gets connected to ground when you switch on PC. I suppose PSU wake up time could be issue when using 2 different PSU's so maybe best to stick 2 identical which should fire up in similar amount of time.

That's precisely what I did with the first dual PSU rig I built, although on high-end PSUs generally all wires are black so you have to work from the pin numbers instead. I prefer pluggable solutions though and it's easier/cheaper to buy an adaptor than to source and make plugs.

[ainarssems]

This is hypothetical but rig with 4 RX480 cards if you can get them or 2 GTX1080Ti cards should cost about £1500 and return over £200 in profit every month and if buying lot of parts probably land a decent discount as well. Obviously there is a lot of risk involved and uncertainty but I could get £15k loan for 4 years paying around £330 a month, make 10 rigs and get over £2k/month from mining so theoretically I could quit my day job ( or night job as it may be in my case). I had a £20k loan for 4 years which I took out to pay deposit for house and it was not too tight paying it back so even if it all goes down south I will be able to afford to pay off loan based on past experience. And in any way kit will be at least worth 50% of original price if it did not work out and I had to pack up mining. So it seems like risk might be well worthwhile. At 10kW+ it would provide enough heating for the house as well making additional savings just need to figure out how to distribute it around house without noise. Our house used to have hot air heating in the past (long before we moved in) and part of the channels are still in the place so that would be an option. I don't think I could go over 10kW which is 40A as I only have 60A supply to house and need to leave spare capacity to run other things so going beyond that would require supply upgrade or move to/rent other premises.

I am not filling in loan application just yet but definitely something to think about for long term and once I have proved I am actually getting money.

If I was to go large scale I would also probably need to think about switching to Linux or cram as much as possible GPU's on one system to avoid or reduce Windows licencing costs. I mean if you cheat and get a free copy or 2 of Windows 10 for disabled to use at home nobody will probably care if they find out who but if I get ten and use it to generate money I guess it could get ugly if get caught.

[ainarssems]

The trick with the RX480 I found, is to lower the GPU speed by about 15-20%. At around that speed, the power consumption drops by about 50W with very little impact on performance.

Still early day but a few options I tried it seems that if I spend £3 less on power per month I loose £3 in income per month, I guess it's not bad considering that hardware runs at lower temps so should be more reliable/stable and last longer.

[ainarssems]

Short term I could get GPU for my HP Microserver as it always runs anyway and while it generates a little money running mining on CPU it's not much. It has PCIe slot but power supply is only 200W and does not have PCIe power connector and does not have enough space inside. But I could get PCIe riser and second PSU to run another GPU outside of case. I am think just for variety and spreading risk somewhat maybe and nVidia card, probably GTX1080Ti.

[ainarssems]

While I am still on the roll, the bios mods seem to promise 25-31Mh/s for RX480 dependant on claimant while reducing power usage compared to 22-24 that I am getting now so might be worthwhile investigating.

[thebluewhale]

I've been following this thread with interest so started researching. It looks promising but I found this:

Bitcoins per Block – Each time a mathematical problem is solved, a constant amount of Bitcoins are created. The number of Bitcoins generated per block starts at 50 and is halved every 210,000 blocks (about four years). The current number of Bitcoins awarded per block is 25. However soon enough the block halving will occur and the reward will be downgraded to only 12.5 Bitcoins.
Bitcoin Difficulty – Since the Bitcoin network is designed to produce a constant amount of Bitcoins every 10 minutes, the difficulty of solving the mathematical problems has to increase in order to adjust to the network’s Hash Rate increase. Basically this means that the more miners that join, the harder it gets to actually mine Bitcoins.


which to me looks as though from the next block halving you're on diminishing returns for the same processor/GPU power.


Or am I reading this wrong?

[thebluewhale]

Originally Posted by ainarssems

I don't think I could go over 10kW which is 40A as I only have 60A supply to house and need to leave spare capacity to run other things so going beyond that would require supply upgrade or move to/rent other premises.

I wouldn't worry about it. The HRC fuses at the supply point have a fusing factor of at least 1.5 times the rating and you are entitled to an 80A supply as standard anyway. So if it blows your friendly local electricity company will come and fix it for you and will probably fit whatever they have 'in the van' which could be anything from a 60-100A

[ainarssems]

Well I though I will drop some early seeds in to my other half's mind if i indeed decide to go that way so that it is easier to convince her later. To my surprise she just said why haven't you done it yet if the numbers are right? So I went on explaining about the risks and that in worst case scenario we might end up paying £330/ month for 4 years for nothing. Well her answer was - who does not take risk does not drink champagne an she had not even had any champagne or any other drink for that matter at the time unlike me And she pointed out that we were paying over £400 for the loan to pay deposit on house and it was fine so £330 would be plain sailing.

I am still a bit vary, I think I need at least to see some money come back to me until I am committed.

[moltuae]

Originally Posted by ainarssems

This is hypothetical but rig with 4 RX480 cards if you can get them or 2 GTX1080Ti cards should cost about £1500 and return over £200 in profit every month and if buying lot of parts probably land a decent discount as well. Obviously there is a lot of risk involved and uncertainty but I could get £15k loan for 4 years paying around £330 a month, make 10 rigs and get over £2k/month from mining so theoretically I could quit my day job ( or night job as it may be in my case). I had a £20k loan for 4 years which I took out to pay deposit for house and it was not too tight paying it back so even if it all goes down south I will be able to afford to pay off loan based on past experience. And in any way kit will be at least worth 50% of original price if it did not work out and I had to pack up mining. So it seems like risk might be well worthwhile. At 10kW+ it would provide enough heating for the house as well making additional savings just need to figure out how to distribute it around house without noise. Our house used to have hot air heating in the past (long before we moved in) and part of the channels are still in the place so that would be an option. I don't think I could go over 10kW which is 40A as I only have 60A supply to house and need to leave spare capacity to run other things so going beyond that would require supply upgrade or move to/rent other premises.

I am not filling in loan application just yet but definitely something to think about for long term and once I have proved I am actually getting money.

Had lots of similar thoughts and discussions with my customer (who I'm helping to build mining rigs for). In their case, they own a couple of large retail stores, the second one they purchased just a few years ago. When we did the maths, we worked out that a fraction of what they bought their second business for invested in mining could provide returns greater than both businesses put together. When you consider also that they estimate their second business won't be into profit for another 8-10 years, mining to them seems lucrative and relatively risk free by comparison. You do have to be realistic and consider the risks, although the same goes for any business venture.

I'm not sure I'd recommend getting a loan to by the hardware though. You may have difficulty getting a loan too if you're honest about its purpose. Perhaps you could scale up gradually instead, building something like 1 or 2 rigs a month. And rather than a loan, you could use a card and transfer the balance to a different card, one that gives interest free credit for a year or two, to give you chance to pay it off before you start paying interest.

And that's something else to consider: Should you use what you earn to live on or pay off loans, or should you re-invest? Bear in mind also that some places, like Aria PC accept Bitcoin payments so you can use what you earn to build more rigs rather than getting into debt. There's also the question of whether you should keep what you earn in cryptocurrencies, potentially adding to the risks, or convert it to fiat currency (ie sell it). Personally I think there's a much greater chance that prices of most of the established cryptocurrencies will continue to climb rather than fall to zero, but it's probably safest to hedge your bets by investing in a few different cryptocurrencies AND selling some of what you earn too.



This is something I posted in a similar discussion I was having on a tech forum a while ago in response to a "Ok, so what's the catch" type of question ...

Originally Posted by Me

There aren't any catches (hence the reason everywhere is presently sold out of most of the popular mining graphics cards) but, as with any business or investment, there are risks to consider when investing in mining hardware:

Long-term mining profitability:
There are a number of reasons it could become less profitable to mine, such as a fall in cryptocurrency prices, an increase in hardware prices or electricity costs, an increase in 'difficulty' or a decrease in mining 'rewards'. Mining rewards don't change too often and usually by the time they do the cryptocurrency's price has increased substantially to more than compensate for the change. Bitcoin's reward 'halvening' happens every 4 years. We had one a year ago, so the next halvening is in approximately 3 year's time. Difficulty is a self regulating mechanism that ensures that new blocks can't be found/mined too quickly. The more mining power that comes online, the harder it gets to mine. However, usually increases in difficulty go hand-in-hand with cryptocurrency price, in that the more difficult it gets to mine, the more the cryptocurrency becomes worth (due to scarcity and higher asking prices).

Long-term mining availability:
Not all coins can be mined (ie Proof Of Work). Some use Proof Of Stake (or even more innovative ideas like Proof Of Storage). Ethereum should be moving to Proof Of Stake in the near future and it's possible Bitcoin and others may eventually do the same, though considering how long it is taking to reach consensus on a fix for Bitcoin's scaling issue and that miner's hold most of the voting power, I think it's unlikely Bitcoin will become a non-mineable cryptocurrency any time soon. And there will probably always be some altcoin worth mining.

Hardware costs and depreciation:
Presently, popular mining graphics cards hold their prices very well, often selling used for as much (sometimes even more) than new cards. That of course could quickly change if it becomes impractical or unprofitable to mine for any reason. However, since the ROI is generally less than 3 months, the risk is probably minimal. Also any risk can be mitigated somewhat by using non-mining-specific hardware, such as graphics cards that are popular for gaming.

The 'bursting bubble':
Some media outlets like to predict a dramatic collapse in cryptocurrency value. These are usually the same ill-informed pro-establishment organisations that show their pitiful lack of understanding by likening Bitcoin to a Ponzi-scheme or predicting the death of Bitcoin (see the growing list of Bitcoin Obituaries published over the last 6 or 7 years -- does it remind you of anything?). If you follow cryptocurrency developments and the more factual news outlets, you'll see the immense growth and potential of a technology that is already very widespread yet only just getting started. However, if you prefer to err on the side of caution, you can of course minimise any risk of losing funds due to falling cryptocurrency prices by immediately exchanging anything earned into your local fiat currency. Considering the state of most of the world economies right now though, I think the wiser bet would be to maintain a diversified portfolio of a number of different cryptocurrencies (properly secured in your own offline wallet of course).


There are probably a few other 'risks' from a business plan point of view, but that's about all I can think of right now. Certainly if you view mining as a business venture (like my customer has), the risks are fewer than most other business ventures. And it would be hard to find another business investment that promises such a quick ROI.

Originally Posted by thebluewhale

I've been following this thread with interest so started researching. It looks promising but I found this:

Bitcoins per Block – Each time a mathematical problem is solved, a constant amount of Bitcoins are created. The number of Bitcoins generated per block starts at 50 and is halved every 210,000 blocks (about four years). The current number of Bitcoins awarded per block is 25. However soon enough the block halving will occur and the reward will be downgraded to only 12.5 Bitcoins.
Bitcoin Difficulty – Since the Bitcoin network is designed to produce a constant amount of Bitcoins every 10 minutes, the difficulty of solving the mathematical problems has to increase in order to adjust to the network’s Hash Rate increase. Basically this means that the more miners that join, the harder it gets to actually mine Bitcoins.


which to me looks as though from the next block halving you're on diminishing returns for the same processor/GPU power.


Or am I reading this wrong?

That's correct, except we're already at a 12.5 BTC reward. The 'halvening' happens approximately every 4 years. The last one was about a year ago so the next one is not until around June 2020: http://www.thehalvening.com/

Like ainarssems says though, the price of Bitcoin tends to increase as a result of any reduction in reward. That's not guaranteed of course but generally it obeys the laws of supply and demand. If it becomes less profitable to mine and miners stop mining as a result, less people are selling and the supply diminishes, driving up the price.

Originally Posted by Markr

Anyone looked at the price / performance of the GPU instances in AWS?

Certainly worth looking into but I think it's unlikely it would be profitable and it may even violate Amazon's T & Cs. I know that some of the Virtual/Cloud Server service providers who offer free trails, specifically forbid the use of their Virtual Servers for mining purposes.