A foray into GPU mining. A humbe noob with a few questions on setup.

Joined
Jan 13, 2022
Messages
5
TLDR. With $100,000 at your disposal, would GPU mining ETH or another coin after PoS and buying RTX 3090's be worth it? Or would ASIC S19 110TH be better and better off sticking to mining BTC at this point?. And if I do go the GPU route, would STOCKX website be reputable for sourcing RTX 3090s at $2300 a piece?

I have done some reading that 6 RTX 3090s can mine as much BTC as a top tier 110TH ASIC miner coming in at half the cost and way less energy consumption.My power cost here iss .07kwh

I have wanted to build a large set of rigs for a long time and have experience building PCs. I'm trying to get an investors pitch completed this weekend because my partners are interested.

If you had $100,000 to start an operation, what would you build and what coin would you mine.

I'm wondering with 6 3090's per rig, what motherboard should I buy, what CPU should I buy, what memory would be best, and would 2 1200w PSU's be enough per rig?

I've read alot of conflicting information online about how you can't mine bitcoin with GPU'S. (But can mine ETH or other coins on Nicehash and have paid out in BTC).

ANY information or advice in regards to this would be very appreciated and 'm hoping this hasn't already been covered.
 
Last edited:
At this moment you should buy 100k in eth.

Mining businesses are rarely successful as they don't quickly move and adapt to a highly volatile market and are unable to source good deals (normally).

I used to build rigs for companies and the only ones who made money were the ones who throw money at it with no hesitation and let me find decent deals and throw rigs together.

Also let it be know you will by mining eth on gpus. Making a pitch and saying you will mine btc on them is extremely foolish and uninformed.
 
At this moment you should buy 100k in eth.

Mining businesses are rarely successful as they don't quickly move and adapt to a highly volatile market and are unable to source good deals (normally).

I used to build rigs for companies and the only ones who made money were the ones who throw money at it with no hesitation and let me find decent deals and throw rigs together.

Also let it be know you will by mining eth on gpus. Making a pitch and saying you will mine btc on them is extremely foolish and uninformed.
Thank you for the reply! But yeah you are correct it would be misleading but I was more or less referring to mining ETH or another coin and getting it paid out in BTC on Nicehash which would be covered in the pitch. I have thought about if all this is not worth it, to just dollar cost average into ETH over the next year or two. I found that sourcing the cards would be (one of) the hardest parts and then we have CMP 22HX coming out (god knows when) which would be way better than the rtx 3090's but I doubt i'd be able to source those when they come out without paying a huge premium.

But if you honestly don't think it's worth doing anymore, especially with ETH PoS coming out eventually, then I'll take your word for it.
 
Last edited:
Spend the 100k on crypto, not cards.
Thanks for the honesty. With way less headache too. I just set my friend up yesterday to dollar cost average $36,500 a year with daily buys into ETH. I'm thinking this is the way to go as well. Just sucks because I've been wanting to get into mining for years but didn't have the capital until this year.
 
God i'm such a noob for not being specific but I was talking about mining ETH and having it paid out in BTC on Nicehash.
Eh, just mine eth at that point. I'd say pretty risky dropping 100k in eth mining. Maybe eth2 doesn't happen in the summer, but if it does i think GPU mining completely collapses. Too much hashing power and the market cap can't grow fast enough.
 
If you want to mine ETH at this point I would also consider ASICs. Apparently these guys are legit and their ASICs are around 4x as efficient as the most efficient GPUs, so about 8x as efficient as a 3090. After ETH 2.0 (lol) you would still be able to mine eth classic with them so they wouldn't be totally useless. That said, their payoff time is > 6 months currently so if ETH 2.0 hits in summer and destroys profits you might be in trouble.
https://www.jinglemining.com/

I think it's risky right now to invest heavily into GPU mining. Profits are down and GPU payoff for most of what's readily available is around a year or more. If you can find something that will pay for itself within 6 months I'd say go for it, but it's hard to find $100k of those deals. As far as GPU choices go, I don't like 3090s because they're inefficient and you don't even get a lot of hashrate despite that inefficiency. For example, a few vega cards will get you more hashrate for less money while being inefficient. I would look into something like the CMP 170HX for $3k instead of a 3090 for $2.6k. At least that way you get high efficiency while still keeping total GPU card numbers low.

I would consider trying a mixture of GPUs and ASICs of various algorithms if I were in your shoes. Then you can focus on buying good deals whenever you see them and avoid paying a premium on things. You also wouldn't lose everything if any single coin tanked.
https://www.asicminervalue.com/efficiency/sha-256

The latest bitcoin miners aren't bad IMO but the payoff time is really long and who knows how long this market will last. Long-term I don't think you'd lose money on these but if the market tanks maybe your payoff time stretches to something crazy like 5 years instead.
 
Split it between top tier level 1 coins that stake well. Stake them and sit back and try to relax.
 
I just spent a couple hours the other night playing with a vega 56 on my test bench, and it runs damn near 1:1 for efficiency with a 3090. They're really underrated cards IMO, but they also aren't "plug and play" like a 3090 where you set your core/mem clock and lock the fan at 80-100%. Each one generally requires a little tweaking of mem timings and voltages for varying silicon lottery.

My Gigabyte 3090: 121mh/s @ 321w (calculated from 303w software reading + 94% PSU efficiency loss) or 2.65w per MH
My Red devil v56: 50.92mh/s @ 141w (wall reading) or 2.77w per MH

Yeah, I easily get 50MH/s on vega56 GPUs but they require memory tweaks and vega64's require BIOS flashing for best results. Memory temps are easier to manage though than on a 3090 at least. That's definitely another consideration. When you have 5-10 cards it's not so bad having GPUs that require more tweaking to get right. But if you have 200-300 GPUs then even something like a BIOS flash adds up to a lot of extra time when you buy new cards / resell old cards. I think I'd rather run amdmemtweak on a vega56 than repad thermal pads on 3090s though.
 
Skip mining, with $100k you could just buy into the market, hit some good Alt coins and 10x your money through 2022. or jjst buy $100k in BTC and sit on it until it rides up, then cash it out, sit on it and buy the dip in 2023
 
Skip mining, with $100k you could just buy into the market, hit some good Alt coins and 10x your money through 2022. or jjst buy $100k in BTC and sit on it until it rides up, then cash it out, sit on it and buy the dip in 2023
Last
 
I was going to make a similar thread but not about 100k, talking like less money for cards to mine but I'm trying to figure out what cards to try and get based on hash rates across many coins to power usage. I notice on whattomine, looks like AMD can't mine certain coins but I'm not sure if that is true or maybe those coins don't matter?

It seems the best bang for the buck is 6600xts followed by 3060ti (even LHRs) but then what happens after Eths goes belly up to miners? No one seems to know for sure so if we keep mining after POS, will these cards still perform very well or not?

I want to expand my mining to make BTC to actually use as cash like a second job. Electricity is cheap where I live so I can do very well with some small mining rigs.
 
This is a terrible time to get into mining. Equipment prices are absurdly high, BTC is down and possibly heading lower, and Ethereum is going away by June 30th. Less than 6 months.

If you had $100,000 that you wanted to put into mining, I would stay away from ASICs. The ROI on those boxes is currently roughly equal to their useful lifespan, meaning that you'd be lucky to get your money out of them before increases in mining difficulty made the boxes unprofitable to operate.

If you wanted to run with GPU's, I would suggest 3090's. These are the only 30 series cards that are not crippled with LHR Low Hash Rate drivers and circuitry. The only way to buy those cards in quantity today is through scalpers, and you'll be paying well in excess of $3,000 USD per card. At a current $6 per card earnings rate, your ROI per card is 500 days, not counting power. And that assumes BTC doesn't drop any lower. Realistically, your ROI will probably be around 18 months. You'll need a good 300 watts per card, and about 150 watts for the board; Figure on 2,000 watts for a 6 card rig. This is more power than a typical home circuit can deliver; I would recommend a 1200 watt PSU to run 3 cards and the MB, and a 1,000 watt PSU for the other 3 cards, running off a separate circuit in your home. When buying your PSU's, keep in mind that you'll need 3 PCIe plugs per card, for a total of 9 PCIe plugs per PSU. Most PSU's don't have this, but a few do. Shop carefully. The alternative is a couple reconditioned 1200 watt HP server PSU's with an appropriate breakout board and cabling, but I have no experience in this approach.

I say again... this is a terrible time to get into mining. You could easily lose your entire investment. Profits are more than a year in the future, and could be several years. Three years from now, your total reward might prove to be ownership of a whole bunch of dusty, highly used, 3 year old video cards. The time to get into bitcoin mining was two years ago, or more. By the time the general media is talking about something 'good', it's usually too late; and the media has been talking about bitcoin mining for at least a year. Not only has the fat lady sung, she's at home, taking off her makeup. It's way, way too late.
Thanks for the reply. Good information here. I'm thinking of dollar cost averaging $36500 a year with daily purchases into ETH over the next couple of years. This seems like the way to go and way less of a headache or risk.
 
Back
Top