GPU MINING For DUMMIES: A complete guide on GPU mining


The following guide is a comprehensive tutorial on GPU mining from planning through execution. It will cover not just the technical setup but the decision making and process regarding profitability and long term gains.

Table of Contents:

1) Planning
2) Hardware & Equipment
3) Environmental Considerations
4) Assembly
5) OS & Mining software
6) Wallet Storage
7) Conclusion

1.1 What to mine
1.2 Electrical costs
1.3 Profitability & ROI
1.4 Up-time
1.5 Monitoring & Alerts
1.6 Scalability

1.1 What to mine
Not all GPU’s are created equal. Some GPU’s mine certain algorithms/coins better then others. For example, NVIDIA GPU’s are better at mining Zcash while AMD GPU’s are more successful mining Ethereum. It is important to decide what you want to mine before purchasing or sourcing your GPU’s. Currently Ethereum is the most profitable coin to mine… however at some point it will go from Proof of Work (mining) to Proof of Stake. At this point those mining Ethereum will need to move to another algorithm.

The good part about GPU’s is that unlike ASIC miners, they can mine any algorithm (although some more successful then others). So switching gears is not as difficult as one might think and we should see future coins using the Proof of Work model.

As of the writing of this guide (Sept 2017) Zcash or Ethereum is the way to go. There are also multi mining pools and platforms such as that will pool your GPU Hashing power and automatically switch to the most profitable coin. Here are some resources to help you make an informed decision:

Reddit, Steemit, and of course the BitCoin Pub mining section are all great resources to ask questions for up to date awnsers. mining is not static! it is evolving monthly. So if you are unsure post a question!

1.2 Electrical Costs
When it comes to mining, there is more to consider then simply how much coin you can mine in a given period. GPU hardware running 24/7 consumes quite a bit of power. If you dont factor in your electrical costs you may actually be losing money when your utility costs monthly are higher then the amount of coin you are mining.

Power costs fluctuate DRASTICALLY depending on where you are in the world. In some areas, it may not be profitable at all to GPU mine. I live on the U.S. East coast where power costs are pretty inexpensive… 0.10 KW/h. You can find your power costs on your utilities bill or by calling your electric company.

Pro Tip! In many areas electrical costs are negotiable! In the united states many states have deregulated energy and you can by power through a re-seller at a discounted rate. Before mining look into energy reseller alternative in your area.

To determine if GPU mining will be profitable, enter your power costs and hash rates into this calculator:

1.3 Profitability & ROI
When mining, there are 2 important factor. How long will it take for you to get a return on investment (ROI) and how profitable is your mining operation. It is important to be conservative when doing these calculations as your GPU mining rig will become less profitable over time as the mining difficulty increases.

Your ROI is very simple. How Many days will you need to mine to reach profits that equal your initial hardware investments. Here is a basic example.

Hardware Costs: $3000
Total Hashing power: 230 MH/s
Total Power Consumption: 800W
Daily profit: $12.25

$3000/12.25 = 245 Days to ROI*

Additionally, These GPU’s we use to mine are extremely powerful and great for gaming. These cards have excellent resale value and there is a large market for them. After 12-24 months you can conservatively resell the cards for 20-50% of their retail value to help offset the difficult increase and claim your ROI.

Profitability is the second factor and there are 2 considerations. How much coin you mine and the interest/market growth of that coin.

You obviously will want to mine as much coin as possible, which we’ve covered in the “what to mine” section. But we also need to consider the long term growth of the coin we mine. Some mining pools will allow you to be paid out in BitCoin. This may be helpful if you plan to invest your profits in to alt coins.

You also want to be smart with how often you take your mining payouts. you may be charged fees per payout or fees for payouts under a certain threshold. You want to avoid transaction fees cutting into your profits.

Best practice would be to read all the guidlines on payouts and fees from the mining pool you are using and to take payouts infrequently… either monthly or quarterly. You may find you will be much more profitable taking payouts less frequently and you may see more growth with your profits coming out in BTC or investing in an altcoin you believe in (NEO, OMG, TenX, Monero, etc…)

1.4 Up-time
When it comes to mining, NOTHING is more important then up-time. Every minute your mining rig is offline you are losing money. So it is CRUCIAL to build redundancy into your design as well as invest in materials and testing to repair your rig if/when something fails.

As with any 24/7 operation, you’ll need to weed out single points of failure. Some considerations are:
-GPU failure
-PSU failure
-Motherboard failure
-Peripheral failure
-OS corruption
-Power outages
-Internet outages

GPU Failure: The most common failure in a mining rig is hardware. By design, mining rigs have multiple GPU’s, so we don’t need to worry about redundancy there as you will already have 5-10 GPU’s per rig. It is important however to thoroughly review the warranty and RMA process for your GPU hardware and contact manufacturer support so that when the time comes you know the steps and the process to get your GPU replaced.

MotherBoard Failure Motherboards are durable equipment without many/any moving parts that typically either fail out of the box or are durable for the lifetime of the board. That said, if your motherboard does fail, you are going to be losing money every second it is offline.

Another consideration in larger mining rigs (6 or more GPU’s) is running all your GPU’s off of one motherboard. I think it is wise to shy away from motherboards with 10-18 PCIE slots. if your motherboard failed and you were running 16 GPU’s on it your ENTIRE mining operation would go offline. where as if you had 2 motherboards if your board failed you would still be able to mine half your GPUs while the other is being swapped out.

Motherboards are stocked in many computer stores and readily available with next day shipping from most online resellers such as Amazon and NewEgg. However you want to be able to quickly swap out a bad board to limit mining down time. with the relatively low cost of motherboards ($50-80 in most cases) it is wise to invest in a second motherboard.

Pro Tip! Be sure to always test your backup hardware. a backup motherboard does you know good if its DoA/a lemon and you realize your mining rig will still be down. so test that backup hardware!

Peripheral failure: Like with motherboards, you dont want your mining operation to go down because your hard drive failed, PCIE risers failed, or you have a bad cable or RAM. These peripherals are inexpensive and you should plan to have a spare on hand of each peripheral. PCIE risers in general are susceptible to failure overtime. For less then $60 you should be able to have a replacement of all your peripherals.

As a wise man once told me… “It is better to be looking at it, then to be looking for it”

OS Corruption: It is not uncommon, especially in Windows builds, for the OS to become corrupted or other software/OS issues to can cause downtime in your mining rig.

There is a simple way to combat OS corruption. Backup or clone your hard drive. Once you have your rig fully configured and mining either clone your hard drive or make a backup image. there are plenty of free or low cost options for cloning or imaging software such as Acronis True Image and EaseToDo Cloning software. I personally like to clone my drive to a spare drive which protects against both OS corruption AND hard drive failure.

It is also a great practice to either record your GPU/hardware/software settings or if your drivers allow it export the settings or export a profile so you can easily restore it if needed in the future.

Power & Internet outages: Power and internet outages are mostly out of our control. but there are steps we can take to minimize the impact. It is strongly recommend to have your mining rig plugged into an uninterruptible power supply(UPS). A small APC UPS can be purchased for as little as $30… even cheaper if purchased through secondary markets.

While the UPS will NOT keep your mining rig up during a long power outage, it will afford you the time to gracefully power down your rig so that an unexpected power outage does not cause a surge or damage to your GPU hardware.

Internet outages will also halt your mining operation. It is a good idea to have a secondary cheap router on hand in the event your router fails. a router can be purchased on the secondary market for as little as $5 or $10.

While you CAN get a redundant secondary ISP, it is not worth the cost as this will incur more monthly recurring costs cutting into your profits and ISP up-time is generally in the 97-99% range. Instead it may be worth while to have a second site/location you can bring your rig too during an ISP outage or disaster (a flood or fire at your current location for example).

1.5 Monitoring & Alerts Having built in redundancy and a good disaster recovery plan are great steps to ensure your mining rig stays up and online. But all that planning wont help if you dont realize your mining rig is down.

It is important to setup email or text alerts anytime your rig is down or stops mining. Most mining pools have a built in function that you can configure to send you a text or email alert anytime the pool detects you stopped mining.

Likewise you can should set alerts with in windows, your UPS, any any other hardware capable of denting outages so you can quickly resolve the issue. Please use the forum to help with any questions on configuring hardware, software, and service alerts.

1.6 Scalability When planning your mining rig, it is important to consider how you may want to scale the operation in the future. building the infrastructure of your rig in a modular way allows you to scale the build more easily as well as upgrade components over time if necessary.

Building a stackable rig for future implementation is important when discussing scaling. Designing a Bottom shelf with clearance and airflow for example is important if you plan to stack a second rig on top of the first one.

Another important consideration is your space. do you have the space to grow your operation physically? do you have the appropriate cable lengths and network infrastructure to grow the rig? do you have the proper cooling environment to handle a large scale operation? This planning should be done prior to the purchase, buildout and implementation of any rig.

Pro-Tip: Check out @ImaginaryPi post on electrical requirements:

Help me pick a GPU
AMD Radeon RX VEGA 64 Mining Rig - INCOMING!
Nvidia 1080 ti Extra Power Inputs (Solved)
Get on the train before it leaves the station!
Noob question about connecting gpu to mb
New To Crypto Looking For Info On Passive Income
Anyone mining in the uk 🇬🇧
The official MasterNet VEGA! RX VEGA Mining Rig FOR SALE!
Congrats to JasonMasterNET for earning the SO AMAZE BADGE :doge:!
Mining & Hardware
Great Links NEEDED!
Mining (newbie) how to?
Zero Energy Costs! To Mine or Not to Mine?
Help get my mining rig turned on :)
Which GPU would you pick to mine Bitcoin?
Getting Started @ The Pub :beers:
*ANNOUNCEMENT* TheBitcoinPub Mining Pool


2.1 Open Air frame
2.2 GPUs & PSUs
2.3 Motherboard and perifpherals
2.4 Cable Management
2.5 UPS
2.6 Environmental protection and cooling

2.1 Open Air Frame Other then your GPUs, your frame may be the most important thing your purchase and design. Temperature control, cooling, scaling, repairs, and upgrades are all heavily reliant on a well assembled well designed open air frame.

While your rig doesnt need to be pretty or a Picasso-like work of art it does need to be functional and allow for cooling, growth and maintenance. The best open air designs seperate components out, allow for proper air space, and are designed in a way that you can easily get to all component in the event in which they nee to be swapped our or repaired. There are a few considerations here and while there are dozens of great designs I will be presenting one in which I believe covers all the bases,

You’ll want to choose a good material:
-Something that conducts heat well
-Something that is light and can be easily moved
-Something that is cost effective
-Something that isn’t prone to produce dust or particles

Lightweight metals such as aluminum are a great choice. Aluminum tubing and aluminum angle are lightweight, cost effective, available at most hardware stores, and dont produce dust or particles that may get sucked into a fan intake. While aluminum angle is easily assembled, I prefer the modularity of aluminum tubing with connectors. is a great online source for aluminum tubing and connectors. the snap in connectors make it easy for anyone to quickly assemble a sturdy box-type frame. The materials can but cut to custom sizes and is fairly inexpensive. I built a 24 X 18" custom cut 2 tier frame for less then $65 including shipment.


A good open air rig is more then just the box frame. You will want to have 2 tiers. An upper tear for your graphics cards and a lower tear for your motherboard, peripherals and power supply.

The bottom shelf should be sturdy and be braced. this will reduce vibrations which can lead to wear and tear or damage over time. As you can see from the photo above we attached 3 pieces of 1/4" thick aluminum across the bottom of the box frame to help brace the shelf.

For the shelf you want to use something sturdy. We chose a 1/8" piece of stainless steel sheeting. This has the weight to remain sturdy and reduce vibration from the equipment running on it. Since this is our first rig there was no consideration for air flow on the bottom shelf. however if this was a second rig, we would instead use a perforated steel sheet for airflow for the GPUs stacked on the rig below.

You will also want to stabilze your components. Drilling motherboard standoffs into the steel base will allow to attach your motherboard and provide airspace for it so it doesnt come in contact with the steel which could cause a short.

You can also use the Aluminum angle to create a bracket to hold your PSUs in place.

The Top Shelf should have a ledge or flange for the GPU to screw into as well as a rear crossbar to support the back of the video card and hold it into place. To achieve this you can attach a piece of Aluminum angle across the frame

For a full list of framing materials, lengths, and a how-to on assembling the frame please send me a private message or check out my Open Air Frame Assembly Guide [Link to Come] Here is what the finished product looks like:

2.2 GPUs & PSUs GPUs and PSUs are the most important hardware components of your mining rig. for GPUs there are several considerations. We already covered “what to mine” in section one and will now focus on which GPUs are the most reliable and profitable for current and future mining.

There are 3 major considerations:
-The initial investment (Cost of the GPU)
-The Hashrate vs Power consumption Ratio
-The long term mining ability.

Initial Investment How quickly you reach your ROI is directly proportionate to how much you spend for GPUs. While GPUs are still somewhat tough to source at cost/without a market things have settled down to where you can find most cards retailers or through online stores. While one GPU may drastically outperform another, it may make sense cost wise to go with another GPU. for example if you are able to buy TWO GPUs with a 18 MH/s output for $20 each that may be a better option then buying one card with a 35 MH/s rate for $500. So keep initial cost of hardware in mind when making this decision. This will vary quite a bit depending on where you live and what options are available to you at the time.

Hashrate vs Power consumption Ratio

There are a few different philosophies on Hash vs. Power Consumption which are influenced not just by opinion but by area. If your power costs are low (or free) Hashrate become most attractive. Additionally with difficulty going up over time, hashing as many coins as possible in the early stages is also a consideration. if you live somewhere with high power costs, power consumption becomes critical. For this guide we will assume all things are equal.

Below is a chart of viable GPUs from a Hash vs. Power perspective sorted by hashrate (no necessarily by best ratio)

AMD Radeon R9 295X2 - 46 MH/s @ 325 W
AMD Radeon RX VEGA 64 - 40 MH/s @ 165W
NVIDIA GTX 1080 Ti - 32 MH/s @ 125W
NVIDIA GTX 1070 - 27 MH/s @ 105W
AMD RX 580- 22 MH/s @ 75W
AMD RX 480 - 20 MH/s @75W

*** Disclaimer. These rates are subject to change and may vary based on driver and bios configurations. There are tons of resources on hashrates. Please do your own research and check the mining forum for real life results people are getting. ***

I’ve personally tested both the GTX 1070 and AMD RX VEGA. I find these to be the most measurable, cost effective GPUs that provide high hash rates with reasonable power consumption. Please check out my AMD RX VEGA mining rig breakdown here:

The RX Vega and GTX 1070/1080 are both great long term mining options. There may be better alternatives or better options based on your situation (Budget,environment, power costs). Please do your research or consult the forum to help make this decision.

Long term mining ability. Just because a GPU mines well now does not mean it will continue to mine well in the future. There are 2 major considerations when considering long term mining ability. Difficulty and Dag.

Over time as coins are mined the difficulty of the algorithms increase. This is a sort of check and balance put in place on the block chain to ensure the security of the ledger and that no single entity can manipulate the blockchain. For more on this, read the bitcoin white paper or read up on how proof of work functions.

The important thing to keep in mind however, is that the difficulty WILL increase. So you will want as much hash power as you can get. whether that is achieved with scale (many GPUs) or Raw power (GPU hash rate) you want to try and squeeze as much hashrate as you can out of your rig while remaining profitable in respect to power consumption.

The second, and more important consideration is the DAG (Directed Acyclical Graph). The DAG basically maps out the algorithms and instructions. It must be loaded into memory almost like a header. Why is this important? Because the DAG grows at each increment. These increments are called “Epoch” and come in 100 hour increments. At a certain point, a GPU will not be able to fit the DAG into memory… this takes quite some time, however in the near future (at approximately epoch 199 or 200) cards with only 4 GB of GDDR5 RAM will no longer be able to handle the DAG and will not be able to mine Ethereum. Additionally, as we approach Epoch 200, the DAG will consume more and more of that 4 GB of RAM cutting into the hashpower of 4 GB GDDR5 RAM GPUs.

What does all of this mean? In Short: Make sure your GPUs have 8 GB or more of memory.

One consideration of going with the AM RX VEGA was its 8 GB of HBM2 memory. RX 480/580 card models with 4 GB of memory are now flooding the secondary market. if you plan to mine, stay away from those 4 GB models.

PSUs are the other important hardware component. This is not only going to power your rig and keep it online, but also factor into your energy costs. All PSUs are not created equal. Some are much more energy efficient then others. It is also important to understand how much power your need. There are 3 major considerations when choosing a PSU:


If your Rig is going to draw 1000W you dont want a 1000W power supply, you want more juice then that as PSUs are not designed to run on their maximum load 24/7. You can refer to this site to help calculate your power consumption and how big of a PSU you will need:

It is also important to not run all your GPUs in series. You dont want to be running several GPUs on the same PCIE or SATA cable. I like to have once dedicated for every GPU or no more then 2 GPUs on one cable. Please refer to the forum and this Thread on GPU mining electrical needs:

PSUs are typically rated into 4 categories: Bronze, Silver, Gold, and Platinum. Bronze being the least efficient and Platinum being the most efficient. Why is this important? Because the more power you waste/use the less profitable your mining rig will be.

PSUs convert AC current from the wall and convert it to DC power for your rig. during this conversion there will some power loss. This also produces more heat. The higher efficiency power supplies require less AC current to produce the same amount of DC power and produce less heat. (which is also important as many GPUs hash better at lower temperatures). You can Save 5%, 10%, or even 20% on energy costs with a Gold or Platinum PSU.


The last factor is cost. PSUs, especially the gold rated ones, can cost up to $200 or more. The more you spend on a PSU, the more profit you need to make to reach your ROI. And as mentioned in the Redundancy/Single point of failure portion of Section 1 above, you are going to want to have multiple PSUs in your mining rig. So choosing the most cost effective PSU and balancing that vs. the most efficient PSU is an important calculation.

Pro-Tip! Instead of purchasing an expensive high end gaming PSU, purchase an old server PSU. Server PSUs are built to run 24/7, are platinum rated, and are very inexpensive ($15-$35). You can purchase a Pair of 1000-1200W PSUs to power your GPUs and a 225W PSU to power your motherboard all for under $100. sells HP proliant PSUs here:

Important! you will also need a breakout board. Server PSUs attach modularly to server motherboards. In order to get 6 pin PCIE interfaces you will need a breakout board. The X6 breakout board is available on both ebay an parallel miner:

2.3 Motherboard and peripherals Purchasing the right motherboard is very important and it is easy to mistakenly buy something that appears to meet the needs of your mining rig but falls short. For example, a 6 card mining rig will require 6 PCIE slots in order to run all 6 cards. There are plenty of motherboard that come with 6 or more PCIE slots. The problem is a lot of these motherboard are not made to handle 6 GPUs running simultaneously and while they can run 6 PCIE devices at the time they weren’t designed with GPU mining in mind. They were designed to handle a couple of graphics cards in addition to other PCIE devices such as RAID controllers, NIC adapters, Card readers and other PCIE interfaced hardware.

Luckily, with the expansion of GPU mining, hardware manufactures have started making motherboards specifically geared towards GPU mininig, ASRock, BioStar and others make several boards designed to handle 6+ GPUs, and even have “BTC” in the model name.

ASRock H110 Pro BTC+ Motherboard

BioStar TB250 BTC Motherboard

While you don’t need a motherboard marketed as a “BTC” or “Minining” motherboard to run your rig you should do extensive research on whether or not the motherboard you choose can handle the GPUs and has shown to be stable running multiple GPUs 24/7.

Resource: The following is a list of some of the more popular mining motherboards:

As mentioned in section 1 with regards to redundancy and uptime; I would suggest against motherboards with 10+ PCIE slots to run that many GPUs all on one board. Not because the board can not handle it, but because if you do have a motherboard failure you bring your entire operation to a halt.

Now that you have all the main components of your rig mapped out, its time to look into the peripherals, bells, and whistles you will need to connect all the components and optimize your design. At a minimum you will need:

-PCIE Risers
-PCIE/Molex/Sata connectors.
-Hard drive/storage

PCIE Risers are extension cables that connect your GPUs to your motherboards PCIE slots. Remember the reason we are building this open air rig is for cooling and airflow so housing your GPUs right in the motherboard would not allow for that cooling and likely overheat your rig. The risers allow you to secure the GPUs on the upper shelf, where air can flow more freely.

There are alot of PCIE risers on the market and most of them are made cheaply and are the component most prone to failure in your rig. It is recommended that your purchase a few spare risers so you minimize downtime in the event of failure.


PCIE risers come in both powered and non-powered versions. It is recommended to use the powered versions over the non-powered 16 pin ribbon cables. The powered versions require less power from your motherboard and from a durability perspective are going to keep you up and running longer. The powered versions are also required for many rigs as most multi PCIE slot motherboards dont have full 16 pin PCIE slots for of their interfaces.

PCIE/Molex/Sata cables Most PSUs and GPUs will come with cables and connectors. But you may not have enough cables to connect all your GPUs or more important connect them all with individual cables. (Running more then 2 GPUs in series is not recommended and likely to cause issues down the road).

If you go with the above recommended route of using HP server PSUs with breakout boards, you wont have any cables included and will need to purchase several 6 pin to 6+2/8 pin cables/connectors. Overall you will want to do a review of all the cable and interface/adapter needs for your components and make sure you have all the required cables. You will also want a few spares for troubleshooting or replacement and most importantly be sure to get the correct lengths so that your cables reach your hardware and arent being stretched.

Hard drives/Storage GPU mining rigs require minimal storage. You will need a drive to run your operating system and GPU hardware. Windows 10 requires 16 GB of drive space for its installation. with software and formatting you can easily get away with a 64 GB hard drive for a windows OS based mining rig. Linux based rigs can run on even less or even run off a usb stick.

64 GB 2.5" or mini PCIE hard drives can be purchased for as little as $35. As mentioned in section one, buying a second drive and cloning your rig after initial tested stable setup or having a backed up .ISO image ready to go for recovery is strongly recommended.


RAM/Memory While GPU RAM/Memory is crucial to the performance of your mining rig, system RAM/Memory is not critical at all. a 4 GB stick of RAM will be enough to power your rig without any issues. With the cost of RAM being so inexpensive, and for redundancy reasons, it is recommended to buy two 4 GB sticks of RAM, keeping your rig operational should one stick fail. a 2 x 4 GB RAM kit will run you as low as $30.

Processor Like system RAM, mining rigs do not require A lot of processing power from the CPU to operate. In fact, it is more profitable not to run a powerful CPU as it wont increase mining performance but will increase your power costs. Using a Pentium or Celeron processor is going to provide all the processing required while consuming less power then an i3/i5/i7 processor and also save you quite a bit of money in hardware costs. Be sure to keep this in mind when choosing your motherboard to ensure the socket can handle the processor you choose.

2.4 Cable Management
Cable management should be considering for any mining operation, especially larger multi rig operation are those expected to scale overtime. Beyond just satisfying those of us who are OCD when it comes to organization, proper cable management while allow you to more easily access components or swap hardware in and out more efficiently. As we have covered several times already, you are losing money every moment your rig is down. to e able to quickly swap hardware and cabling in and out is an important design consideration.


Some best practices for cable management include:

-Use cables of appropriate length, with enough slack to prevent pulling but not so much that it will coil up and tangle.
-Use velcro in place of tie wraps whenever possible. Velcro allows for easy removal and recabling where ties need to be cut.
-when cutting the excess of a tie wrap, be sure to cut right at the knuckle to avoid sharp edges that could cut wiring
-use your frame to secure and space your cables out evenly and channel them to the appropriate areas.
-Avoid crossing cables across the rig, and design your rig so that components connected with cables are closest to eachother.

2.5 UPS
A UPS (uninterrupted Power Supply) is strong recommended for all mining rigs. A UPS will NOT keep your mining rig running through a power outage… what it will do is allow you to gracefully and safely shut down all the components of your mining rig so that a power outage and sudden shutdown will not damage any of the components.

Protecting your investment is critical. You dont want to be caught setting up an RMA or warranty claim on a GPU that failed during a power outage nor do you want your rig down while you wait 7-10 business days for a new one to arrive. A 1500W power supply can be purchased for as low as $65. Most UPS also have the ability to set up email or text alerts that will notify you of an outage which is a nice feature to have.

2.6 Environmental protection and cooling
Once you have your rig set up and running it is important to protect your rig against its surroundings. Depending on where you store your rig, you may need to worry about things such as dust, debris, moisture,and other elements. Some considerations and failsafes to consider are:

-to keep your rig in a cool, dry area.

-keep your rig elevated off the ground in the even of a spill or flood.whether on a shelf or even building your frame out with legs to elevate the shelves.

-To protect against dust and debris you can line the frame of your rig screen door screen which will allow air flow while cutting out dust and particles. Covering fan intakes with pantyhose is an old IT hack that many people still employ that also keeps dust and particles out.

-Use compressed air to blow any dust or debris out of your rig on a periodic basis.

In addition to keeping your rig dry and clear of debris, proper cooling is also important especially in larger scale multi rig mining operations. While the subject of cooling is vast and beyond the scope of this guide it is important to know that keeping your rig cool is critical not only to keep it from overheating or causing damage but a lot of high end GPUs will also have increased hash rates at lower temperatures.

Channeling heat away from your rig with duct or tubing, fans, or vents can make a considerable difference in how cool your rig will run. While air conditioning is a very effect way to keep your rig cool, be sure to consider the power costs of running air conditioning 24/7, as this may cut into your overall profits or even put you at a loss depending on the power consumption of the air conditioning and electrical costs in your area.

Welp, building my miner - and I forgot to buy the power supply

3.1 Where to store your rig
3.2 Power requirements
3.3 Building to Scale

3.1 Where to store your rig
Choosing where to store your rig is important as the environment will be integral in ensuring that our operation stays online and is protected against hazards. Like any electrical equipment, it is best to keep your rig in a cool dry place to prevent against moisture damage or overheating. You’ll want to store your rig somewhere with:

-Ample space for airflow
-Good ventilation
-temperature control
-free of dust and debris

You’ll also want to keep your rig elevate off the floor to protect against any floods and potentially enclose your case to protect against debirs. This will be covered in more detail in the hardware and assembly sections of this guide.

A finished basement or temperature controlled hardwood or laminate floor room with ventilation make for good choices for small scale operations. Sever rooms, or warehouses if available are also obviously well equipped to house mining rigs. Your grandma’s attack or small IT closet are probably not going to be safe for longterm use.

3.2 Power Requirements
For this section, I refer to @ImaginaryPi Guide for power and electrical needs:

The complete guide can be found in the Mining & Hardware support forum:

3.3 Building to Scale
When considering your environment, it is important to plan things for future growth if you intend to scale the operation. The needs and requirements to run 25-50+ GPUs is drastically different then running just 5 or 6 GPUs.

Your electrical infrastructure, storage, cooling, and potentially even zoning and permit requirements will need to be planned in accordance with how large your operation may grow. It is recommend to consult with an electrical engineer and potentially even your town or city for guidelines on scaling your operation to something of commercial size and power needs.




4.1 Frame assembly
4.2 Prepping for components
4.3 Component installations
4.4 Cabling
4.5 Disassembly and testing
4.6 Customization

4.1 Frame Assembly
"Measure 10 times, cut once" -Someone really smart

This old proverb needs to be running on loop in your head the entire time you assemble your frame. Spacing out your frame to properly hold your GPUs is really important. You need to make sure your GPU measurements as close to exact as possible when ordering/cutting your material lengths. If you are using the above recommended tubing and connectors from “Esto Connectors” you need to be sure to factor in the 3/4" added to the length by the connectors.

It may also be worthwhile to make a full scale model out of scrap wood or cheap plastic to make sure you get the measurements right. When building my first rig, i used some left over 1 x 1 pieces of wood left over in my garage to build the top shelf of my rack to get an exact measurement for the spacing on my GPUs. Once you have your prototype it is easy to get the exact material lengths and purchase or cut those materials. From there it is just some screws and/or elbow grease to get your frame assembled


Once assembled, verify once more that your components and GPUs will fit properly. The last thing you want to is to start cutting/installing shelves and brackets to find out you made a cutting error and it wont all fit right.

4.2 Prepping for components
While most of the hard work is measuring and building the initial frame, there is prep work still needed to properly install your hardware components. You’ll need to consider the following:

-Screw holes will be needed to attach the front of your GPUs to the frame and hold them in place
-A rear support will be need to support the weight of the GPUs
-Braces and a bottom shelf will need to be installed to hold your motherboard and power supplys
-motherboard standoffs will be needed to secure the motherboard and rise it above the shelf
-angle brackets are a great way to secure PSUs into place

Before drilling the screw holes for your GPUs, space them out and mark where they will all go to ensure even spacing and proper airfflow. Using a cardboard template of the width of your GPU can be helpful in this instance. Once marked drill the holes to the width of the screws you are using (i used #6 machine screws). After drilling each hole, secure the GPU to test fit and angle. then remove the GPU and move on to the remaining screw holes.

While some designs will use the frame itself to support the back of the cards, you may want to install something a bit more modular that can be more easily adjusted. a piece of aluminum angle installed horizontally across the top tier is a great fit to support the rear of the card while maintaining some flexibility should end up swapping out GPUs or add additional GPUs with a different form factor.

Some designs will support the GPUs by using shelves with PCIE risers secured to that the GPUs plug directly into. This is a really nice design however it does require a bit more building material and by eliminating a shelf under the cards you allow for closer to complete 360 degree air flow.

Now that we have our upper tier designed and assembled, we need to work on the lower tier which will house or motherboard and power supply components. As mentioned in the above sections, minimizing vibrations of these components helps prevent wear overtime as well as keep your sanity by eliminating the constant noise.

In our design, we used 3 pieces of 1/8" thick aluminum to brace the the lower shelf:


For the lower shelf itself, using a heavy metal, like stainless steel can help keep the shelf secure and free of vibration. We used a 1’8" thick piece of stainless steel sheet. You can secure the sheet to the frame with self tapping or machien screws, but the snug fit and weight alone was tight enough and vibration free as is.


With our shelf installed, we need to create the ifrastucture to secure our Motherboard and supplies. A short piece of aluminum angle spaced out to the width of your PSU is a good choice to secure it in place. We used industrial strength adhesive to attach our angle,


Last we used 1/4" CPU standoffs (female on both ends) as a base to hold our motherboard in place. Carefully mark your bottom shelf where the 4 screw holes for the standoffs will need to be drilled in the bottom shelf. From there simply drill the holes and attach the standoffs from the bottom with screws. This will allow you to install and secure your motherboard later on

With your frame now assembled, you can move on to installing the components.

4.3 Component Installations

With all the hard work and measurements out of the way, its time to have some fun and start putting your rig together. It is helpful to get all of your components out and accounted for. You’ll need your GPUs:

Look at those beautiful beast Vegas

Your PCIE Risers:

Motherboard and PC components:

Power Supplies: (2 server PSUs and a PSU for the motherboard components):

As we have already lined up and tested the positioning on the GPUs, We will move on to assembling the bottom tier components. Installing the motherboard screwing it to the standoffs and positioning the PSUs making sure the brackets hold them in place securely:

Once you’ve secured everything in place, you are ready for cabling.

4.4 Cabling
Now that everything is secured to the frame its time to begin installing our cabling. It is important to make sure all cabling is proper lengths, reaches without stretching and without too much slack which can cause tangles and make it difficult to troubleshoot later.

Attach your cables, one at a time moving inside out. this well prevent your cables from crossing between each other. Use velcro to attached your cabling to the frame or any fasteners you have installed. Once you got all your cabling complete, you are ready for some testing.

4.5 Disassembly and testing
Before you beggining your OS and software installation its important to do some testing and to disassemble your rig Making sure your rig POSTs and your bios is recognizing all your GPUs is an important step before finalizing your physical installation and moving your rig to its final destination. Once it POSTs and sees all your hardware, you are in a good spot to proceed forward. if it doesn’t you can easily troubleshoot the issue while everything is easy to get to.


You will next want to know ahead of time how to take everything apart and reassemble it so that when you need to troubleshoot or make repairs down the road you’ve weeded out any issue that prevent from quickly swapping out components. This will also allow you to make any last minute customization or bells and whistels.

4.6 Customization
With our rig ready to go into OS/software installation and the physical assembly completed and tested its time for any last minute customization, tweaks, bells, whistles, and branding. A $2 can of red spray paint for instance can go a long way to make your rack stand out a bit from the cold drab wire racks you’ll be sitting it in for the next year,

For our build we gave all the aluminum and steel a coat of red so our frame matched the black and red theme of the AMD RX Vegas that the rig would power.



5.1 Operating System
5.2 Drivers and firmware
5.3 Mining software
5.4 Tweaking and testing

5.1 Operating System
Mining software can be run on multiple platforms and operating systems. Windows 7/8/10 are all supported by most mining software and pools. In addition several linux distobutions also have support including customize OS with the mining software built into it such as EthOS.

While it is beyond the scope of this guide to cover the installation of all operating systems we will cover some of the considerations on choosing an operating system and best practices for installation.

A windows installation or stand alone Linux install will allow for more flexibility then running say EthOS as you can install several different mining softwares and switch based on profitability and stability. However if you just plan to mine one coin, using something like EthOS is likely to be more stable. The profitability really will not fluctuate much and this is more of a user preference, although i do personally prefer a windows installation due to all the reporting and alerting i can do and find it easier to tweak drivers settings within windows OS.

Prior to installing your operating system, its best to update the firmware of your BIOS, as doing this after a windows installation may cause some issues or you may have to re-do any settings or configurations. once complete, you can move forward to the OS install.

5.2 Drivers & Firmware
Once you’ve chosen and installed your operating system, it is important to update all of your drivers, firmware, and get the most up to date versions of software that support the hardware you are using. While most hard ware will work “out of the box” with OS drivers, up to date versions will be more stable and allow for more tweaking.

In our build, going from the default VEGA drivers to the updated AMD mining drivers saw a 20% increase in hashrate and considerable reduction in power consumption. This drastically increase the profitability of our rig.

You may want to steer clear of alpha/beta or non confirmed drivers and software unless they are the only ones that provide support for the hardware you are using. these pre-stable release versions can cause crashes and downtime for your rig.

Once you have a working set of drivers, and a stable environment it is recommended to do a backup of all settings and/or do a full image backup of your hard drive. You dont want to be repeating this timely process again in the event of a system or hardware failure.

5.3 Mining software
Once your rig is setup with an operating system and optimized with up to date drivers you are ready to install your mining software and start mining! There are several algorithms and coins to mine with GPUs as mentioned earlier in this guide… and with that several different options for mining software. There are tons of great resources on installation and setup of mining software. the most popular/profitable are:

Claymore dual miner (Ethereum + [coin])


NiceHash (multiple coins, chooses most profitable, BTC payouts

MultiMiner (multiple coin support)

Resource youtube, reddit, steemit, and right here in the Bitcoin Pub there are great resources for setting up and tweaking your mining software as well as choosing the most profitable option based on your hardware.

Pro-Tip! Before you even begin purchasing your GPUs an mining hardware, install and test out the different mining software on your computer to see how they install and function. It is helpful to be familiar with them so that you hit the ground running when its time to install them on your rig.

5.4 Tweaking & Testing
Once you have chosen and installed the software you are going to use it is CRITICAL that you test your software and HashRates with different settings. You can see DRASTIC changes in what your HashRate is in certain software based on your driver and bios settings. For example, simply changing the fan speed on your GPU can result in +/- 5 in hash rate.

Take the time to do your due diligence. this is one area where holding your rig back from mining is profitable in the long run. test your hash-rate/power consumption of your rig on every potential setting. When setting up our rig we literally adjusted every setting 1 unit at a time and retested the rate/power after each reconfiguration. The result? A HashRate:PowerConsumption ratio so much better then what others had been seeing and reporting that people have called us out as liars until we confirmed with the screenshots and settings.

We took a rig originally operating at 38 Mh/s | 355W per card to one hashing at 39.5 MH/s | 160W per card. We also found an alternative configuration of 43.5 Mh/s | 245 W per card. When you put that into perspective, that testing that took a week probably shaved a month off our ROI and increased our profitability by about 12-17% daily.

Pro-Tip! Purchase an energy/power usage monitor. These are inexpensive and allow you to determine how much power your rig is using, how much your are spending daily/monthly on electricity and send power alerts if you reach certain thresholds or go offline. You can also use it to test the power consumption of your rig at different settings. We use one from Belkin - Wemo



6.1 Taking Payouts
6.2 Crypto vs. Fiat
6.3 Storing Payouts

6.1 Taking Payouts
Once your rig is up, stable,configured for your mining pool, and actively mining it is time to start thinking about what to do with your payouts. How and when you take your payouts will be a big factor in how much profit you make and how quickly you reach your ROI.

Most mining software and pools will pay you out in the coin that you are mining. So for instance if you are mining ethereum, the pool will pay you out in ethereum. There are some pools and services however such as nice-hash which will pay you out in BitCoin or other currencies. So depending on how you plan to use your mined currency, you may want to consider one option vs. the other.

If you plan to convert all of your mined zCASH to alt coins, joining nicehash so you are paid out in BTC which can be easily traded with alt coins may be an attractive option for you. conversely if you plan on keeping all of your mined zCASH as zCASH you are better off being paid out in zCASH.


Another consideration is how often you take payouts. while it may seem attractive to get paid out as frequently as possible, the less often you take payouts the fewer transaction fees you will end up paying. which over 12-24 months can add up. Additionally some mining pools/software will charge a higher fee for payouts less then a certain amount (i.e. less then 0.1 ETH). You want to avoid these fees whenever you can. Monthly or quarterly payouts are recommended unless you are either really nervous about your mined coin living on the pool for that long or believe you can make more then the transaction fees trading your mining profits for alt coins.

There is no right or wrong way to take payouts depending on your investment profitability and long term strategy on mining profits (holding them in the mined coin vs. trading). But it is smart to calculate the costs and risks of taking more infrequent payouts with less fees vs. the alternatives.

6.2 Crypto vs. Fiat
Prior to reaching your ROI, it may be intuitive to cash out all of your mining payouts/profits to US dollar/Fiat currency to pay back towards your ROI. depending on your circumstances this may be the case. for instance if you bought some of your hardware on 6 or 12 month 0% credit offers and need to pay that back prior to getting hit with interest on the purchase, then applying that profit to Fiat/US dollar would make a lot of sense.


Alternatively, the amount of profit you make could be drastically increased by keeping your mined profits in Crypto and earning compounded interest on that coin as the value of ethereum or other mined coins increases in time. Those who mine, and are involved in the crypto community typically project the market to increase over the next 12-24 months. If there isnt debt to be paid, keeping your mined profit in crypto is likely to increase your profits and help reach your ROI quicker.

With any investment decision, doing your own research and factoring your own philosophies and risks is of the utmost importance.

6.3 Storing Payouts
While you may already have a wallet for the coin you mine, it makes a lot of sense to create a separate wallet just for your mined coins. This will allow you to more easily tack how much coin you have mined and run any tools or reports regarding profit and projects to ROI.

For instance, if you have an exsisting wallet with 10.43 Ethereum coins in it, having your mined payouts go to that same wallet may make it difficult to track how much Ethereum you mine over the long haul, especially if you cash out to trade some of that Ethereum.

Resource There are several tools, apps, API integrations, and excel spreadsheets available right here in the bitcoin pub to help track your investments and profits:



In conclusion, GPU mining can still be a profitable endeavor and also a great project to enjoy and learn about crypto and how Proof of Work functions.

It is important to understand that where you live and your energy costs will determine whether GPU mining is profitable for you or how profitable it may be.

Redundancy is extremely important to any mining rig, and building in that redundancy to reduce downtime is crucial to avoid losses during hardware failure.

Taking advantage of any free resources to help reduce ROI is crucial for anyone entering the GPU mining world. if you cans source PC components, building materials, software licensing or even run your rig somewhere with free power costs, your time to ROI will be drastically decreased.

Lean on the community and The BitCoin pub for information and guidance. We have one of the most active and helpful communities right here in CryptoNation!

For any questions, comments, or corrections regarding this guide or on GPU mining in general feel free to reach out to me via private message or email me at:


Awesome stuff. This post should get shitloads of traffic.



Good article. Hopefully, others can contribute some troubleshooting. My rig seems to reboot perodically (Thermal reasons IDK???). It’s only got two RX 480s in it. It mines 45Mh/s, but sometimes after it reboots it will only pull 25Mh/s and then reboots stop. I’ve used ethminer and EthOS, both will do it. It’s maddening.


It would be helpful if you could provide more details:

-what are you using for motherboard/power/risers/etc
-card manufactorer?
-what are you using for mining software?

are there any errors when it restarts? do you get the generic “windows has recovered from a critical failure” or does it bluescreen? have you set the bios not to restart on system failure (this would allow us to see the bluescreen error"

in these scenarios most common is overheating or driver issue


Thanks for the shout out. Yeah

  • Running on a MSI B250 PC MATE motherboard. EVGA supernova 650 Power Supply. No risers in the equation at the moment (But it happens with the risers in play too.)
  • One card is a PowerColor Radeon RX480 8Gb. The other is XFX RX480 8Gb. The cards look identiacal except one has a back plate and the other does not.
  • Currently Running Ubuntu 16.04 LTS, I’ve used EthOS as well. Driver is 1.2 AMD-APP
  • Mining software is ethminer

No errors just reboots. I configured Ubuntu to restart mining on startup to keep things going. But like I said, sometimes after a reboot it will mine at a much lower rate for no apparent reason. If I could throw a limit into the command, it might fix it? Who knows? Just curious if others see this happening as well? Any ideas?


What is the draw on your power supply? do you have meter to see how much you are using? you could potentially be seeing spikes beyond what your power supply can handle.

Do you have the same issue with just 1 card? or is it stable on 1 card? if it IS stable on 1 card is that true for both the XFX and the power color or is just stable for one of them?

if it is stable for either card individually but no both, its most likely a power or heat thing.

how hot are the cards running? do you have anything to cool with? if you crank the air conditioning in the room do the crashes stop?

the rig hashing different after a reboot isnt all that suprising. it could be either your card settings being reset after failure or a safety mechanism throttling the card.

Does your mining software have any logs? have you checked the ubuntu system logs?

step 1 should be determining if the problem happens with just 1 card or isolating the issue to 1 of the 2 cards


I have a Kill A Watt. Will that help me determine the draw on the PS? Or just the whole system? Maybe it’s the same.

Yes, it runs 100% fine on just one card. Doesn’t matter which card either.

I do have a separate personal barrel fan blowing down the cards.

I see no temps on my rig sadly. On EthOS it did, and on ETHOS it ran the fans at nearly full tilt and yet it would still reboot.

I’ll see if I can find some logs. Also, let me know if the Kill A Watt is really going to help here. I bought it to calculate energy use, but didn’t think it would help with troubleshooting.


For the Kill-A-Watt its the same… your power supply is basically the entire system. thats what powers everything. What is it showing for power consumption?

The fact that it runs fine on one card… and it doesnt matter what card it is makes me pretty sure its a power or heat issue. your 650W supply may power 1 card no issue but not both. likewise 1 card may not produce enough heat to overheat but both cards might (especially if the rig doesnt have great air flow)

The fans blowing full speed are an indication the cards are running hot… fans not running full speed is actually a GOOD sign.

the kill-a-watt should help us alot. it will tell us how much power the rig is using and if your running to close to your PSU’s limits. also if you can see the power consumption as it reboots, that could tell us alot as well.

if you have a 650W power supply thats using 600-650W continuous you are certain to see crashes.

id start with getting those power readouts and maybe jacking up the AC or setting up tons of large window type fans to try and cool the unit.

your bios/motherboard should have temp readings. if it doesnt you can get a seperate sensor. you can also try running your rig on windows, if anything ti will give you more t work with for troubleshooting


Also posted on Steemit. feel free to upvote :slight_smile:


Ok, so I moved the rig where air flow should be better. I suspected that could be a contributing factor.

I have the Kill-A-Watt in line with it right now and it seems to be settling in at 347W use. I’ve not seen it spike over 350W in the last 20 minutes or so. I’ll see if it reboots in this configuration.


350W should cause no issues. thats with 2 cards?

if so then id say overheating was more likely the issue. your doing the right things now, lets see how it behaves.


Yeah, 350W running both cards. The thing got about 30 minutes of mining in and reboot!

Should I put a card on a riser to separate them?


how hot is it running… like if you touch the card or put your hand near it is it super hot?

its worth putting them on risers, getting them a ton of airspace and seeing what happens. do you habe air conditioning in that or big fans you can plug in to cool them?

also, what motherboard are you using? id think 2 GPUs should be fine for any motherboard but who knows.