BUYERS GUIDE - How to build your own GPU mining rig - What to buy!



Don’t use laptops to mine, they do not manage heat well. I’ve seen the wiring for the fans melt.

If the case of the mac pros have poor factory ventilation, we may be able to take the shell of it off. All experimental though. Until we try it out, we won’t know.


Mining on a mac is more like playing with a toy, but it’s indeed the very first experience I had with mining:

It was fun and I got to try out a few things here and there, but it’s also impractical. For one, the macbook gets very hot. I wouldn’t want to run it for very long. Secondly, churns so much resources, everything else becomes non-responsive, esp. for the GPU miner. Surprisingly, the CPU miner didn’t bog everything down anything like the GPU miner would.

I did most of my playing on the laptop, but I did also briefly tried out my Mac Pro from 2008, but it’s graphics card wasn’t any better than the laptop. However, you can definitely fit a powerful card in and get going. In which case, Homebrew, XCode, and git are good tools to have installed so you can build things from source.



Nice! :+1:t4: Thanks for getting me heading in the right direction… I’ll definitely look into the Graphics Card’s you mentioned. Sounds like this may be a good place to get me started in the mining world.

I’ll post progress once i’ve started. Any more info that you guys can share please do… Enjoying learning this stuff! :sunglasses:



Full disclosure did not read the whole post, but at what time line wise does it make financial sense to build your own rig instead of using a cloud mining service like hashflare? Let’s say mining BTC only. Thanks in advance


Based on what @mwlang has said about cloud mining and what I’ve read, it never makes financial sense to to cloud mine. There seems to be a mixture of good and bad reviews on cloud mining in the topic Who’s investing in CLOUD MINING companies?

As far as building your own rig, I would say it financially makes sense when you feel confident that the ends justify the means. Meaning, when you have calculated the expenses of purchasing the equipment, creating the necessary infrastructure, and calculated your return on investment. Do tons and tons of research until you feel confident you can earn some extra bitcoin. It’s completely your choice to mine and I believe that if you start mining, you will become addicted. Even when it’s not profitable you are supporting the network thus securing the future of your investment.

You aren’t going to be able to mine BTC with this GPU setup, at least not for a profit. Bitcoin mining is done with AISC machines which, as far as I know, aren’t a DIY type machine. You can set your GPU mining rig to mine for Nicehash where you mine what they determine is the most profitable algorithm and pay you in BTC.

Guide to GPU Mining Profitability
BUILDER'S GUIDE - How to assemble your GPU mining rig!

For a power supply, this i believe is by FAR the best option:

  1. HP Prolant platinum rated power supply ($15-35):

  2. Breakout board to 6-pin connectors ($19):

I’m using these in my rig and they work great. Much more efficient the the gold evga gaming power supply we started with,

my recommendation:

2 x 1000 W power supply (for your GPU’s)
1 x 225 W power supply (for your motherboard)
3 breakout boards

When your mining rig stops, your losing money. so you want to build in redundancy. put your motherboard/hard drives on a small PSU seperate from the GPU (this makes troubleshooting easier).

Split your GPU’s between two 1000W (or 1200 or 1500) supplies. this way if one PSU goes you lose half your rig instead of all of it.

Also, these PSU’s are so cheap and they break out boards are modular that swapping out a bad one takes minutes and the cost is low ($15-$35)

In short… for a fraction of the cost you can use more power efficient power supplies that were built to run 24x7 that can easily be swapped out in the event of failure extremely quickly.


That is awesome… but for me it is a REALLY bad idea.

you now how a single point of failure for 19 GPU’s.

id much rather have 2-3 motherboard where if one fails I only lose 5-10 cards instead of 19 cards.

redundancy and weeding out single poitns of failure are extremely critical for anything that needs to run 24/7, especially things that are losing money when offline


I’d like to know why people seem to prefer the PCI 6 slot over the Molex power connection for PCI risers. I’ve read that the Molex are not good. Need more data!!!


The molded ones can potentially catch fire:


Brother Peter,

Geography aside, I saw your comment on the ASICs miner top, I just liked it, and I think the ASIC’s are better than the old GPU’s because they are so much faster. Also, even if they just run one algorithm, that algorithm can be used to mine various different coins if one becomes unprofitable. The way I looked at it, I threw $3,000 from my NEO profits into buying two D3 miners because the return on one year ($100,000 at the time, now down to $30,000 as the difficulty has gone up) seemed much better than what one BTC would do in a year. I also looked at it as like a secure hedge investment because I was new to altcoin trading and I was a little skeptical as to if I could continue to turn such profits. With all of the tools you have shared with B90X, now I’m sure I can continue to make profitable investments and trades, so thanks for that!

Anyway, that is why I think us cryptonauts should compile our tech knowledge and efforts to start MANUFACTURING ASIC miners. Tigerdirect does not sell ASIC chips last time I checked, but Alibaba sells them wholesale. However, I wouldn’t know what to do with the chips. You are the tech guy, I mean, I know you studied IT, but maybe I should get in touch with an electrical engineering friend of mine who introduced me to a few of the first bitcoin exchanges I used, to see if he is interested or know hows. Perhaps after my first two D3 mining rigs arrive in the next few months I can try to replicate them. I don’t really want to take them apart because it voids the warranty, but it would be cool to have a lot more of these mining rigs.

Can you share with me what kind of rigs you are using, and why? I mean, GPU’s are the old tech, so I’m really surprised to hear you on YouTube saying you have $50,000 worth of rigs up and mining now and that they are profitable. Props for that, though :slight_smile: . I suggest stepping it up to ASICs ASAP, but like you say, maybe there is something I don’t know or understand, hehe.

Hope to hear from you good sir,

John Dailey
Sole Proprietor of Geographic Consulting International

Racks over Water


woops, I meant for that to be a DM, I’m such a n00b ><


Hey @FloridaWater! Thanks for being here. I think you touch on one of our concerns with the ASIC miners, the quick decline in profitability. With ASIC you are locked into one algorithm. Yes there are multiple coins but that is still 1 algorithm. With GPUs we have multiple algorithms to choose from and it’s very simple to point each rigs to whichever coin we wish to mine. Not to mention, when we have our own coin, we will be able to have immediate mining support. Check out some of our other concerns with ASIC rigs here:

Concerns about ordering Antminers from Bitmain


You know…looking back on history, I nearly got into Bitcoin and mining 4 or 5 years ago and at that time, I was looking at these ASIC miners that were starting to come out from Bitmain. There was a lot of questions raised at the time whether people preordering 2 or 3 months ahead of shipment time would ever see those machines and once they got them if they’d still be profitable. Every one of those concerns are still concerns today. Then I wondered, surely these guys have, by now, gotten their production issues out of the way and can sell these things at a rapid clip and JITM (just in time manufacturing) to boot. But it’s still the case that an open order waits 2, 3, more months to ship.

So I was researching buying the machines direct on (the Chinese version of Aliexpress) and ran across a comment about one of the sales item that explained that “this isn’t the real deal from Bitmain. Besides, if it were, Bitmain isn’t shipping you their latest until after they’ve mined with the hardware for at least 6 months themselves.”

More or less as I’m paraphrasing and it was a Google Chinese => English translation…

Anyway, I haven’t really seen proof this is really what’s happening with Bitmain miners, but I haven’t exactly seen anything to say “no, way!” either. Makes you wonder!

Plus, how freakin’ hard is it to encode a software algo into a processing chip? How the heck is Bitmain still the one and only ASIC manufacturer for Bitcoin miners after all these years?


Hey man. You didn’t know?

Why would the sole producer of such machines give them out immediately to public before they have fully reached their own internal capacity?

Not a dig, I just assumed that people knew this.

It’s simple Chinese economics.


Why would I know? :-o

I mean, 5 or so years ago when I thought about getting into it, I thought it was a silly thing waiting 2 or 3 months for a multi-thousand dollar purchase and dropped out of cryptosphere altogether.

Now, I’ve only had skin in the game a mere two months.

Assume nothing, brother. Assume nothing! You must P90X us all the way through. :sunglasses:


Sorry. Didn’t want to come off snarky.

My bad bro!!!

Yeh. This has been out on bitcoin talk years ago. Sometimes old news gets forgotten.

If I remember correctly, even at one point there was a thread with a seemingly legit individual confirming this.

Invidia and AMD don’t have the same internal goals to mine. They build a new chip. Get it out ASAP to public.

Bitmain business model is different. They RELY on their own product. They need to milk it before they give it to their customers.


Nothing snarky there! I know what it’s like to have so much knowledge, one loses sight of what newbies don’t know. That’s the way it is with me and IT with my clients. I even get to the point where I mistakenly treat new clients as having at least the basic knowledge of my ongoing clients, so I really have to take time to go through the finer details with new clients.

Reflecting on this, I just now realized it takes about a full-year of “education” before we are operating on the level that makes the relationship and the workflow truly fluid. It’s something I’ve always kind of knew, as in, “new clients are hard!” but it just now clicked why. It’s that time to get to know one another, get to know their business model, get to a common and terse vocabulary that is unambiguous, and it’s the flip-side as well as the client gets to know all about me and how I roll.


Yes, let us manufacture ASICs mining rigs here in the U.S.A.! :smiley:

I wanted to start manufacturing solar panels a few years ago too but it was nearly impossible for me to beat the economies of scale with mass production of solar panels in China. However, these mining rigs are different in that there are already backorders.

I am going to continue to network with people who have experience in computer and electrical engineering in an attempt to create an American knock-off of Bitmain. :slight_smile:



great post! I made a list to do some further research thanks for putting this together.