Cryptic Crypto- How can I make $5 per day efficiently?

[JoeHam]

Like tens of thousands of other folks who viewed Wills latest video on crypto it got me thinking. I noticed a mention of NiceHash by him and thought of starting there.

In a few minutes, I set up an account and put to work a computer that is not being used during the week. The computer is a Hades Canyon NUC with an i7-8705G CPU running at 3.1 GHz and a Radeon RX Vega M GL GPU with 4 Gb of RAM. The machine has 8Gb of RAM and runs Windows 10.

So far it has been in continuous employ for 36 hours and I have made 0.25 USD. Slightly underwhelming but given my minimal efforts so far and extremely limited understanding I am not completely surprised. IF I am going to pursue this it looks like a purpose-built rig is the way to go.

After more hours of research I want to set $5/day as a goal and build something to accomplish that.

I want to mine, or get paid, in Bitcoin due to a number of reasons.

My current understanding is that the GPU is the main driver for success and likely building up to a multi GPU open frame system is the best with 6 GPU's as a reasonable long-term goal but the GPU's are crazy expensive currently.

Will has a good starting list of components on his video.

That makes me think of building a system with scalability to 6 GPU's but I may start out with 1 GPU if it will generate a reasonable return.

Can anyone with experience say if 1 GPU can make this happen or do you need multiple GPU's to get a $5/day target return?

PS- My preference is not to use inefficient and loud ASIC miners.

 

[Hedges]

While I have no experience with bitcoin, it is my understanding that complexity increases with time so the amount of bitcoin that can be mined per unit time with given hardware decreases. Value may increase to offset that. But as hardware follows Moore's Law, I would expect dollar return from given hardware to decline. So you need to account for that when estimating payback horizon.

Bitcoin Difficulty Chart - BTC Difficulty

The Bitcoin difficulty chart provides the current Bitcoin difficulty (BTC diff) target as well as a historical data graph visualizing Bitcoin mining difficulty chart values with BTC difficulty adjustments (both increases and decreases) defaulted to today with timeline options of 1 day, 1 week, 1...

www.coinwarz.com

 

[JoeHam]

While I have no experience with bitcoin, it is my understanding that complexity increases with time so the amount of bitcoin that can be mined per unit time with given hardware decreases. Value may increase to offset that. But as hardware follows Moore's Law, I would expect dollar return from given hardware to decline. So you need to account for that when estimating payback horizon.

Yes, and of course the value of Bitcoin fluctuates wildly which prevents a precise solution of course.

To keep this simple I wanted to focus on Bitcoin as stated and use $5/ day as a rough goal estimate.

 

[LB3]

Like tens of thousands of other folks who viewed Wills latest video on crypto it got me thinking. I noticed a mention of NiceHash by him and thought of starting there.

In a few minutes, I set up an account and put to work a computer that is not being used during the week. The computer is a Hades Canyon NUC with an i7-8705G CPU running at 3.1 GHz and a Radeon RX Vega M GL GPU with 4 Gb of RAM. The machine has 8Gb of RAM and runs Windows 10.

So far it has been in continuous employ for 36 hours and I have made 0.25 USD. Slightly underwhelming but given my minimal efforts so far and extremely limited understanding I am not completely surprised. IF I am going to pursue this it looks like a purpose-built rig is the way to go.

After more hours of research I want to set $5/day as a goal and build something to accomplish that.

I want to mine, or get paid, in Bitcoin due to a number of reasons.

My current understanding is that the GPU is the main driver for success and likely building up to a multi GPU open frame system is the best with 6 GPU's as a reasonable long-term goal but the GPU's are crazy expensive currently.

Will has a good starting list of components on his video.

That makes me think of building a system with scalability to 6 GPU's but I may start out with 1 GPU if it will generate a reasonable return.

Can anyone with experience say if 1 GPU can make this happen or do you need multiple GPU's to get a $5/day target return?

PS- My preference is not to use inefficient and loud ASIC miners.

You will never make any money using GPUs for Bitcoin. Other currencies like Etherium, maybe.

The problem with Bitcoin is that the professional miners have invested millions developing custom chips that are designed only for crypto mining so their hash rates and efficiencies dwarf what we can achieve using idle GPUs.

The good news is that China’s crackdown on Crypto mining has resulted in many of these mining rigs becoming available on the open market.

Good luck.

 

[Hedges]

So far it has been in continuous employ for 36 hours and I have made 0.25 USD.

How many kWh were consumed? How many by the GPU?
It may be a net loss unless your power is free.
How does GPU cloud computing compare?

https://lambdalabs.com/service/gpu-cloud

The GPU in my laptop delivers a fairly impressive number of MFlop, but only low-precision numbers. More bits, and its performance is far worse (like hundreds of times slower for twice the bit length, not sure why.) Some GPU models (at much higher prices) are far higher performance. Consider whether the bits precision of yours is appropriate for the job.

I think mine is k5000 or some similar laptop version. Compare to V100 to see drastic difference.

https://www.amazon.com/NVIDIA-Quadro-K5000-Graphics-PNY/dp/B009L8E4UO

https://www.amazon.com/NVIDIA-Tesla-Volta-Accelerator-Graphics/dp/B07JVNHFFX

 

[madsci1016]

Nicehash is a good way to start, it mines whatever is most profitable (and it supports) and pays you in bitcoin. You loose a little in their "service fee" more or less but that's the cost for making it easy.

Put in your GPUs into their estimator for profit projections: https://www.nicehash.com/profitability-calculator/

One thing it's easy to miss, even with nicehash you still have to "tune" your GPUs for efficiency. like tuning an engine. Boost memory clock till it becomes unstable and back off a bit, reduce GPU clock till it looses hash speed and bump it back a bit. Reduce power limit till you loose hash speed and bump it back a bit. Use MSI Afterburner for that. Example:

(note One of the nicehash programs does do this for you. see here: https://www.nicehash.com/blog/post/nvidia-and-amd-graphics-card-oc-settings-for-mining)




To answer your questions in the other thread:

Are you mining with a GPU (non-ASIC) based rig? GPUs

How many GPU's ? 2x 3060Tis LHRs and 1x 1070.

What currency ? Ergo right now because it's hot. Will drop back the ethhash on nicehash when it pops.

I flip everything to bitcoin in an exchange and use bitpay to buy gift cards (mostly for groceries and home depot) to use the earnings without the tax man seeing.

 

[LB3]

On a side note, I work with some brilliant people. One friend suggested building a mining rig back when BTC broke $1.00. We decided that for an optimal ROI on our budget, we should target a miner pulling about 8kw.

We were going to run it at my house and agreed to split the power bill for this device. And agreed on a split that would cover my increased air conditioning costs. My wife nixed the deal worried about us getting stuck with an extra $600 per month in power bills (and the extra scrutiny from cops looking for grow-ops).

In the end we abandoned the plan. I could have been a billionaire. -That’s my “if only” story for today.

 

[JoeHam]

You will never make any money using GPUs for Bitcoin. Other currencies like Etherium, maybe.

I would be happy mining Etherium and get paid in Bitcoin.

I was hoping $5/day was an achievable goal. No interest in competing with the big farms.

 

[JoeHam]

How many kWh were consumed? How many by the GPU?
It may be a net loss unless your power is free.

I consider my solar setup a "sunk cost" so will ignore that for this exercise.

For the sake of keeping this a focused discussion I am just trying to figure out what would be required in just the rig hardware to accomplish $5/day production.

 

[madsci1016]

BTW if you don't currently have GPUs i have bad news....

World sucks for getting GPUs right now. You have to either pay 3x-4x their usual price or spend weeks hunting a deal only to loose to bots who place orders faster than a human.

 

[JoeHam]

Thanks @madsci1016 for the answers to my questions posted in the other thread.

This greatly helps me focus my efforts to climb the learning curve.

Here is the other, somewhat poorly focused thread for anyone who is curious :

Open letter to Will Prowse

Will = "Right now I am using my driveway array power to run GPU and ASIC crypto miners, and I will probably add more solar, so I can make more money faster. I am trying to push as much money out of this system as possible. It is very addicting to see how much money I can make in a day with pure...

diysolarforum.com

 

[JoeHam]

BTW if you don't currently have GPUs i have bad news....

World sucks for getting GPUs right now. You have to either pay 3x-4x their usual price or spend weeks hunting a deal only to loose to bots who place orders faster than a human.

Yes, if making money is the only goal then I think trading GPU's would be far more profitable.

MSRP is just a starting point and stock selling out in less than 10 seconds thanks to the bots seems to be common.

And I thought getting raw LiFePO4 cells from China was difficult!

 

[madsci1016]

Yes, if making money is the only goal then I think trading GPU's would be far more profitable.

MSRP is just a starting point and stock selling out in less than 10 seconds thanks to the bots seems to be common.

And I thought getting raw LiFePO4 cells from China was difficult!

You joke but I'm living it. A retired miner gave me a box full of old and broken GPUs early last year to play with and I shelfed it. Sold them all this year on ebay for more profit than I'm every going to make mining. Broken 1000x series going for $200+ each; Working 470s for the same. It's nuts. Paid for my new 3000x series I'm buying even at today's prices. Sold a personal 1060 for $100 more than i originally paid for it 4 years ago. Going to sell a working 1070 once my new 3080 gets delivered.

 

[JoeHam]

Not joking, just finding it amusing as I sit on the outside looking in. Laugh vs cry has always been my way.

I would have done the exact same thing were I in your position.

That will likely prove to have been a very good choice for you if Ethereum abruptly pivots away from proof of work in Q1 or Q2 of next year as they have stated.

That's part of the reason why IF I pursue this I would likely build a 6 GPU slot rig and only populate 1 or 2 spaces to get started.

 

[madsci1016]

Yeah i don't know enough to claim to be a good forecaster at this but with what i have read; my betting odds is Ethereum drops the bomb and pivots due to political pressure, the billion dollar mining spaces pivots to altcoins and causes cascade failure of each, and then attempts to liquify their assets flooding the market with cheap GPUs now that they can't mine. And Nvidia goes bankrupt for a while and all GPUs loose their value.

A least thats the outcome i'm planning any of my current GPU acquisitions too. If i'm wrong no harm. If i'm right I'll have at least paid off the GPUs i have now.

 

[JoeHam]

Well, I am curious enough to pursue this experiment and see if I can generate $5/day with a rather limited rig. The following is what I have ordered for the experiment:

Ryzen 5600G CPU which has onboard RX Vega 7 graphics. I will run this alone for a day or two mining on NiceHash to see what the CPU based GPU can do. I am assuming it can beat the 0.25 USD per day the computer I described in the first post was generating.

This will be running on a MSI B550-A Pro motherboard which is NOT a mining MB but will allow me to add a few additional GPU's.

I also got a 256 Gb Lexar NS100 SSD, probably overkill.

In addition, I am using Patriot DDR4 16 GB ram (2x8GB) which is also overkill.

It will be powered by an EVGA 850W Gold series power supply which will allow some room to grow.

In the beginning, I will add one PowerColor Red Devil AMD Radeon RX 6600 XT GPU to see what the daily USD output is for this setup. I could add more of these if it goes well.

Please note that I am far from an expert on this stuff. I had never considered mining until less than a week ago and have been climbing the learning curve in the past 6 days. TBH, I currently know next to nothing but want to do some hands-on learning now.

I should have recieved all the pieces/parts and get the rig assembled early next week and will report back on whether I am getting close to generating my $5/day arbitrary target or not.

I should credit Will for the idea of the RX 6600 XT GPU and the Lexar/Patriot brand names from his crypto video.

 

[Hedges]

Next, you will learn who FinCEN is, as they attempt to determine if your $0.25 or $5 /day is being used to finance terrorism or launder dirty money.

I'm told they have a file on me, but won't disclose what's in it for fear of exposing their investigative methods.
They're also the ones who detected "Client #9"

 

[JoeHam]

In case anyone is curious about the current prices I had to pay:

PowerColor Red Devil AMD Radeon RX 6600 XT GPU $648.45

Ryzen 5600G CPU $258.96

EVGA SuperNova 850W PSU $143.90

MSI B550-A Pro motherboard $129.99

Patriot DD4 3200 mHz 16GB RAM $65.99

Acrylic ATX MB open frame test bench $35.29

Lexar 256 GB SSD $31.59

Sales tax not included.

AMZN prime shipping on all except the GPU which came from eBay and had free shipping.

 

[Surferdudemi]

This is a pretty good source for info on profitability: https://whattomine.com/gpus

The reason to use many GPUs on one MB is that you split the CPU power cost and mobo across all of them. It takes very little resources with a modern CPU to shuttle data back and forth. You don't need a big drive, either. I put one together with an old 80GB SSD I had lying around. Make sure you have a good Internet connection with low latency to avoid rejected shares.

You'll need extenders to add more than a couple of GPUs to most mobos. Example (I have no recommendation): https://www.amazon.com/VER009S-Riser-Cable-Bitcoin-Mining/dp/B093H1JHKS

 

[JoeHam]

The reason to use many GPUs on one MB is that you split the CPU power cost and mobo across all of them.

I completely understand.

The issue is GPU prices which currently approach twice the MSRP.

As designed, the rig is somewhat scalable and built for the possibility of using more GPU's as discussed in post #14.

I'm just not willing to pay those prices at this point, especially with Ethereum going from proof of work to proof of stake, perhaps as early as Q1, 2022.

IF GPU prices return to earth and there is still profitable work to be done with them I can scale up at that time.