Cryptocurrency mining is a profitable business. As we learned from the “miner’s confession“, if there are initial funds and a strategy, even the cost of equipment for mining coins can easily reach several million rubles.
However, if you are just at the beginning of the journey, that’s okay. This article will cover the basics of mining, overclocking and other tricks.
Mining is the provision of computing power to protect a cryptocurrency network with a probabilistic receipt of rewards for the work done by the equipment. Unfortunately, many legislators mistakenly associate coin emission with mining. It is incorrect to say that miners issue cryptocurrency or participate in its emission, because for most modern currencies, leaving up to 10 percent of miners every day will not affect the emission rate by more than 1 percent.
Even if at the end there are only ten miners out of a million, the emission will not stop and will quickly reach the same level, that is, it will adapt in complexity.
Miners provide cryptocurrency protection from possible blocking or canceling transactions. The more miners, the better the defense and the more difficult the attack, up to the impossibility of an attack due to the lack of the required amount of necessary mining equipment in the world.
For this work, any network of the cryptocurrency PoW system (Proof of Work) pays the miners with their coins from the issue of new coins and transaction fees.
If we draw an analogy with the existing banking system, then the cryptocurrency network is the Central Bank (emission), mining pools are commercial banks (packaging of transactions, distribution of awards, competition among themselves), and miners are clerks (they do the main work). Banks cannot work without employees, but this does not mean that they are the issuers of currency.
How to mine cryptocurrencies in 2021?
Who can be called an experienced miner, and how can this experience be counted? There is no definite answer to this question, but there is no doubt that an experienced miner should be able to adapt to constantly changing conditions. Still, the crypto market is very volatile, and coins, the profitability of mining for which today is 20-30 percent lower than the optimal coin for you, can give a double profit tomorrow.
You don’t have to go far for examples. The Cortex coin (CTXC), which gave $ 4.17 per day from one 1080ti card on August 13-15, 2020, gave less than $ 1.8 per day for two weeks after that, and on August 30-31 again soared to $ 3.31 per day from the card.
However, mining itself for most miners is a passive income option, so most do not jump from one currency to another every day, but focus on a stable average income on a large and expensive coin.
However, every miner needs to know the capabilities and limits of their graphics cards.
It would seem why, because there are profitability calculators? But even such a wonderful tool as the 2CryptoCalc currency profitability calculator for different video cards will not be able to tell you how much the profitability of mining will be higher than the increase in utility tariffs, and whether the profitability of the selected coin will decrease five or more times tomorrow. Also, he will not be able to say how well the mining room is ventilated, when the tariffs at the multi-tariff meter are switched, and what level of noise from the equipment is acceptable for you during the day, evening and night.
ASIC and video cards for mining – which one to choose
Initially, it was assumed that everyone can mine at the capacity of the central processor of their computer. However, it quickly became clear that a universal and powerful computer processor loses in power and efficiency to video cards, and they, in turn, lose to even more narrow-minded devices – rare FPGAs and ASICs that are very common now.
Asics (ASIC) are highly specialized mining devices for one algorithm. They are produced by a small number of companies with sharply opportunistic pricing policy, which, moreover, are not averse to memorizing them on new equipment before sending them to the customer.
Mining on ASICs is very simple, and at the same time, very unreliable in terms of profit. Competition is limited solely by the production capacity and greed of the manufacturers of these devices.
Cheap and powerful ASICs quickly replaced video cards from most of the old algorithms, which did not aim to protect against them.
The biggest problem of ASICs is the narrowness and constantly decreasing income in cryptocurrency, associated primarily with the release of new models, whose energy efficiency is many times better than the old ones.
A video card or GPU is a widespread universal device that can work on many new algorithms and has a good residual value (40-60 percent) in the secondary market even after several years of operation. After the expiration of the warranty period, or better before the end, they can be sold to gamers.
Most new cryptocurrencies are trying to be as decentralized as possible and inaccessible for mining on ASIC devices. In addition, the energy efficiency of video cards can be changed many times even within the same series of video cards, while the patterns are different for different algorithms. This makes valuable not only the fact of the presence of the video cards themselves, but also the knowledge that the miner acquires during operation, as well as the miner’s awareness of the situation on the cryptocurrency market.
And the most important thing: with limited power, video cards give much higher profitability per 1 kW of power (several times) compared to ASIC devices.
Therefore, for personal use, miners mainly use video cards, and mainly the latest or penultimate generations.
Mining video cards
Powerful gaming video cards are mainly produced by two major global corporations – AMD and Nvidia. By the color of their logo, miners and gamers refer to them as red and green, respectively.
Mining uses not gaming, but computing capabilities of video cards, while using special OpenCL libraries from AMD and CUDA from Nvidia. This usage for cards is not “non-standard”. In some ways, the work of a card in mining is similar in load to drawing or calculating textures of the same type in games, therefore, with the growth of gamers’ needs for gaming capacities, their calculated power proportionally grows.
Mining on AMD graphics cards
The architecture of AMD graphics cards was originally focused on fewer logical blocks and more stream processors, so it was much better suited for early mining algorithms, successfully performing many small operations. It is not surprising that before the advent of ASICs, it was on AMD cards that more than 95 percent of the mining work was performed, and the rest was done mainly on processors.
Modern mining algorithms for new coins require a good level of core-memory interaction, which is perfectly implemented in game cards, but difficult to implement in ASIC devices.
It so happened historically that the most successful project among new cryptocurrencies with the positioning of an asic-protected algorithm was Ether – Ethereum, ETH – and AMD video cards turned out to be more adapted to its mining than competitor products.
These two facts have shaped the face of mining for years to come. From 2015 to 2018, inclusive, the lion’s share of video cards in mining came from AMD, and in the overwhelming majority of cases they mined – or “dug” in the jargon of experienced miners – Ether.
Basic setup of AMD cards for Ether is quite simple:
- first of all, we determine at what frequency the memory will work as much as possible, with a step of +25 MHz. We find the one on which the video card driver starts to fly – reboot – we retreat from it 50 MHz down into the stable region.
- we find the core voltage at which we are satisfied with the power consumption of the card, the lower the voltage, the less the card “eats”, but the lower the maximum speed. The dependence is disproportionate, usually when consumption decreases by 20-25 percent from the nominal, the speed drops by only 10-15 percent.
- we find the maximum stable core frequency (also at +25 MHz), we retreat from it down to 25 MHz.
A finer tweak will give us 5-6 percent extra speed or 3-4 percent energy efficiency. We will not dwell on it in detail in this article.
Mining on Nvidia video cards
The basic architecture of NVidia graphics cards had significantly fewer stream processors combined into larger CUDA blocks, which made them less efficient in mining until 2016.
Although already in the 9xx series on the Maxwell architecture it was possible to mine ETH at an acceptable speed, but in 2016, cards on the Pascal architecture appeared – the 10xx series – they were the ones who made a turning point in the total dominance of AMD cards in mining.
Nvidia cards have always been appreciated by gamers for their energy efficiency and good cooling performance. They did not warm up like AMD cards and did not drop frequencies in games due to overheating. The green logo of the manufacturer has become a symbol of optimal power consumption, and Nvidia really tried to bypass AMD in this parameter at the release of each new series.
The 10xx series was no exception: in 2016-2018, the cards of this series could not yet surpass AMD in the power of Ethereum mining, but in the level of maximum energy efficiency they confirmed their “green” status. On the 16nm process technology, Nvidia achieved better results than AMD at 14nm — the 4xx and 5xx series.
Important Feature: Nvidia cards show the total GPU consumption, not just the core consumption, so they cannot be directly compared to AMD cards in terms of driver readings.
A correct comparison is possible only by measuring the actual consumption of the entire rig from the outlet. As they often say “from the wall”.
Unfortunately, the 10xx series cards were not designed specifically for mining, so in 2020 the 10xx series miners encountered some speed drop on the Ethash algorithm – the current Ethereum mining algorithm. To maintain speed, power consumption had to be raised, which significantly changed energy efficiency. AMD faced a similar problem at the beginning of 2017 and it was solved by releasing special drivers, and later by including a special “computation mode” in the base drivers. Nvidia solved this problem only in the new series of cards (16xx and 20xx).
Despite the identified problem with Ethereum mining at 10xx, green cards still have a very significant advantage. They turned out to be better tailored for many other mining algorithms, because most of the cryptocurrencies created after 2017 did not want to use existing algorithms, but tried to make a unique product to one degree or another.
Before moving on to a detailed consideration of algorithms for green cards, we will answer an important question, how to assemble a working computer for mining with several video cards – or, in mining jargon, a farm and a rig?
How to build a mining farm – choosing an operating system
We will not dwell on the hardware, because now all modern motherboards support mining on four or more video cards (most of 8 or more video cards using cheap port extenders).
We will be primarily interested in the operation of operating systems. Here the overwhelming majority of miners are clearly divided into two groups – work on Windows and work on Linux, and the vast majority of the latter use not Linux in its pure form, but ready-made mining operating systems like RaveOS.
Benefits of the Rave OS system
Rave OS is a system for miners that allows them to control what is happening with mining farms from a distance. At the same time, the developers promise the stability of mining, high hash rates and minimal downtime, which usually lead to losses.
You can manage three devices for free, and if there are more of them, then each work will cost $ 2 for a month.
The main advantage of the system is its simplicity. To install, just write the image to a medium and connect the latter to the device. After that, the system will automatically determine the hardware model, select the appropriate settings and allow mining. In addition, Rave OS has a large community, which means there are plenty of overclocking combinations.
Features of mining on Windows
The main disadvantages of Windows mining are the following points:
- the need to independently install many necessary mining programs, drivers, disable auto-updates, antiviruses;
- knowledge of the nuances of stable operation and restarting miners;
- lack of a built-in restart algorithm in case of failures – on power supply, on video card freezing, and so on.
It is also necessary to install a remote monitoring or remote desktop system, because the user is not always near the rig. And the rigs often do not have extra monitors, mice and keyboards, since this is an extra expense for the periphery.
An important point when working with Nvidia cards running Windows is that initially the memory of video cards at standard frequencies allows a small number of errors that are not critical when rendering images in games, but can be critical for accurate calculations. Therefore, for the operation of CUDA systems, Nvidia engineers have provided a special mode with a reduced memory frequency (P2).
The transition of cards to this mode is carried out automatically, but miners often want to get the maximum hash rate and seriously increase the memory frequency – including up to values of +700, +1000 and even higher. As a result, two factors are added – a decrease in the frequency by the P2 mode and an increase by the miner, while the miner works successfully, but when the miner is turned off, the P2 mode switches back, and plus the manually set frequency remains. This leads to frequencies that the cards cannot hold and, as a result, the driver crashes when the mining program is turned off or restarted.
The best thing to do is to disable P2 mode for CUDA. This is done through the nvInspector utility, or rather its nvProfileInspector subroutine. We find the CUDA parameter – Force P2 state, turn it off, then start the miner.
Important: This setting is restored when you install any new Nvidia card into the system, do not forget to double-check.
When using MSI Afterburner to monitor the status of cards – a necessary utility for any miner – we will immediately notice an increase in the operating memory frequency during mining, as well as feel an increase in the speed of memory-dependent algorithms like Cortex mining. In this case, the memory frequency will not jump, which means that you can more stably overclock to higher memory frequencies. Moreover, this is not just a higher plus sign in the settings, but the real operating value in MHz.
Mining on NVidia video cards in 2021
As mentioned above, Nvidia video cards are much more versatile in the choice of algorithms, but this does not mean that everywhere they are much more profitable than AMD video cards.
Mining algorithms are divided into “hot” and “cold”. Hot algorithms, as a rule, use the video card core quite strongly, so their performance significantly depends on the core frequency, and the maximum core frequency is reached at high voltages, which proportionally increases the card’s consumption.
Cold algorithms use memory very heavily, so they quickly run into a bottleneck in bandwidth or memory access speed. Thus, it makes no sense to overclock the core very strongly, because this will not give a speed increase. It is easy to achieve maximum efficiency with such algorithms. It is necessary to determine at what core frequency the maximum speed is reached, and then get this core frequency at the lowest possible, but at the same stable core voltage.
Summer 2020 for green cards was not very rich in algorithms. Leadership in profitability was mainly held by the following algorithms (coins): Ethash (ETH – Ether), MTP (XZC – Zcoin) and Cuckaroo30 (CTXC – Cortex).
Learn more about these algorithms for Nvidia cards.
Mining Nvidia on the Ethash algorithm
Until recently, Ether’s algorithm was one of the most efficient for Nvidia 10xx, 16xx and 20xx series cards. However, it is designed in such a way that every 3-4 days the complexity of working with it grows slightly, which requires more and more memory on video cards. At first, he had enough cards with 2 GB of memory on board, but at the end of 2020, it became a prerequisite to have more than 4 GB of memory for video cards so that she could mine ETH.
And here it was discovered that due to the peculiarities of the architecture of the 10xx cards, they cannot effectively mine Ether in the new era. The speed of mining on them dropped immediately by 25 percent in a few months, the speed can be restored by seriously increasing the frequency and voltage of the core, which means seriously reducing the energy efficiency.
Conclusion: after 2020 and beyond, we cannot call ether mining on 10xx cards the most effective alternative.
For younger 16xx and 20xx cards, such a problem is irrelevant, so this really cold algorithm can be successfully used for mining, and a good memory reserve – 6 GB and more – allows you not to worry about the growth of the DAG file for at least a couple of years.
The most efficient configurations are shown below. Here’s the GTX 1660.
And the RTX 2070s.
As you can see, an increase in the frequency and voltage of the core does not give not only a significant, but in general no increase in speed. Therefore, the optimal energy efficiency of this series can be considered the consumption of 220-230 W per 100 Mh.
Overclocking the core to 650 mV on new cards seriously reduces the maximum speed, but does not increase energy efficiency due to the card’s transition to P3 mode with a significant decrease in the memory frequency so necessary for this algorithm.
Let us make a separate reservation that in its development Nvidia is now approaching the number of stream processors to AMD cards, so the new series (30xx and their modifications) have become even more successful in confronting this algorithm.
There is a big “but”. The Ethereum developers want to fundamentally move away from the PoW model, so the profitability of this algorithm may significantly decrease in the next couple of years.
Mining Nvidia on the MTP algorithm
The MTP algorithm was developed in the transition – or even flight – of the Zcoin from the old algorithm, which was actively threatened by ASIC devices.
This algorithm was originally designed to be ASIC resistant. In addition, its feature is the requirement for a video card memory capacity of at least 6 GB, since this parameter does not grow over time, as it happens with the Ether algorithm.
The algorithm is quite hot, it depends very seriously on both the core frequency and the memory frequency. Simply lowering the voltage does not necessarily increase energy efficiency.
As you can see in the screenshots, lowering the core voltage on 1660 cards to 650 mV leads to a significant drop in the hash rate, but does not lead to an increase in energy efficiency. Thus, you need to look for and know the optimal minimum point for your cards, as well as a good maximum point – without overheating. Nevertheless, overheating inevitably leads to a significant degradation of the characteristics of the card and even to its failure.
For 10xx series cards, the efficiency range is from 15 kx per 1 W of power – up to 22 kh / W. For the 20xx series, this range improves from 19kX / W to 31kX / W.
New series were supposed to improve this indicator by a certain value (20-30 percent), but not multiple (as with ether), so even in 2020, 10xx series cards are successfully mined using this algorithm.
For 1080 and 1080ti cards, it is imperative to activate the tablet to get the maximum hash rates – 3.4 Mx and 4.4 Mx, respectively.
Mining Nvidia on the Cuckaroo30 algorithm
The algorithm for the Cortex coin is a hybrid algorithm that uses both the core and memory, but the advantage is still in the memory. It can rightfully be called a very “cold” algorithm.
The algorithm is perfect for mining on video cards with 10 or more gigabytes of memory free. For example, on 1080ti cards, it gives a result comparable to the profitability of the MTP algorithm at maximum settings and Ethash at medium settings. However, the actual power consumption of the cards from the outlet will be 40 percent below the MTP and 25 percent below the air.
Important. The Cuckaroo30 algorithm is one of the algorithms with a non-constant (“ragged”) load. Unfortunately, systems like HiveOS show only momentary load, which will vary within 30 percent. In the miner window (for example, Gminer), of course, you can see the real load, but this information is not displayed in the web interface, and it is with it that most users of this system interact.
For example, for 2070s cards, when the power consumption limit is set to 240 W in RaveOS, such users will not understand that with this algorithm the card will not consume more than 200 W, since the short stage of active kernel loading is replaced by a long stage of working with memory.
This is clearly visible in Windows, where when the maximum PL is set to 111 percent (245 W) in the window of every second consumption check, we see only single values above 220 W in a row. The rest of the time the consumption is significantly lower (minimum 158 W), and the final average value in the program is 197 W. We also note that on a platinum power supply, these power values, according to measurements, practically correspond to the real consumption from the outlet.
On Windows 7 systems, as well as Linux systems (RaveOS, MMP and others), you can use Gminer to start mining on cards with 8 GB of memory.
For 10xx series cards (1070, 1070ti and 1080), this is not very productive (speeds 1.55, 1.6 and 1.65, respectively), but perfectly suits for 2070 and 2070s cards, on which the hashrate almost linearly depends on the set Power Limit value. With PL = 240W, the 2070s 8GB card consumes about 197W and has a hashrate comparable to the 1080ti 11GB.
Features of other mining algorithms
Models of cards with a small amount of memory, just old cards, and some modern models, unfortunately, cannot mine the above algorithms or are less effective in them. Nvidia also has a large number of interesting and profitable algorithms for such cards.
The main ones in terms of profitability in the second half of 2020 are the Beam and Ravencoin coin algorithms – they are called BeamHash-III and Kawpow.
Mining Nvidia on the BeamHash-III algorithm
The third version of the algorithm for the BEAM coin is positioned as an ASIC-resistant algorithm for GPU mining.
Initially, it was believed that it needed more than 3 GB of video memory on the card, but rather quickly the creators of the LolMiner mining program included with their program support for cards with 3 GB of free memory, that is, memory that does not display an image in the Windows system.
When overclocking the core, the algorithm quickly runs into memory in the region of average power consumption values on 10xx series cards (about 70 percent), and a further increase in the core frequency almost does not lead to an increase in the hash rate. Therefore, this algorithm cannot be considered hot, while it gives a good average annual profitability on cards with 3-4 GB of video memory.
Due to the presence of a large number of old cards on the algorithm, it does not appear in the top in terms of profitability for cards with more than 6 GB of video memory on board.
Mining Nvidia on the Kawpow algorithm
The Kawpow algorithm, the main currency of which is RVN, burst into mining quite brightly, but after three months it stopped getting into the top profitability.
It would seem that its rise was short-lived, but it is necessary to note the key advantages: the algorithm works even with cards with 3 GB of memory, it is not capricious (many capricious cards on other algorithms work confidently on kawpow), the speed correlates well with the level of PL cards, and the coin RVN is listed on many major exchanges.
In terms of a combination of factors, this algorithm confidently ranks second in importance and profitability after Ethash.
Overclocking is similar to Beam, let’s not dwell on it. By the way, it was Beam who pulled off part of the 3 GB cards, which made it possible for kawpow miners to feel more stable and confident in 2020.
In addition, after the ETC update and the outflow of a part of the 3 GB cards for this coin, we will more often see Kawpow in the top profitability even for large cards.
Mining Nvidia on the Cuckarooz29 and Cuckatoo32 algorithm
Both algorithms (abbreviated as C29 and C32) are the development of the GRIN cryptocurrency, which was initially positioned as GPU-oriented and asi-resistant, and therefore the developers periodically slightly change the algorithm to make it as difficult as possible to create highly specialized devices.
Due to the rather serious binding of these algorithms to one coin, the popularity of which has slightly decreased, mining of this moment on the C29 algorithm is practically not found in the top profitability.
Unlike its brother, the C32 algorithm has significant requirements for the amount of video memory: at least 8 GB of video memory is required for mining, while, like for the Cortex algorithm, the top cards of the series have a significantly higher hashrate. Therefore, this algorithm often appears in the top profitability for 1080ti, 2080ti, 3060ti, 3070, 3080 and 3090 cards.
The algorithm is not capricious, it works very stably even on cards with weak memory.
Mining Nvidia on Equihash-xxx algorithms
In 2016, when the anonymous ZEC coin appeared on the new equihash algorithm, it was a good challenge to the dominance of ETH in mining. Initially, the coin was mined only on the CPU, but less than a month later, a GPU miner appeared, which gave significantly better results on green cards than on red ones.
The Equihash algorithm is highly dependent on overclocking the kernel and does not have significant intermediate files for placement in the memory of video cards, so it was a relatively easy target for ASIC device designers. Despite the threat of centralization of the network, the creator of the coin did not change the algorithm, which gave rise to a lot of rumors about their affiliation with the largest ASIC miner manufacturer Bitmain.
Unlike the main currency, various clones really felt threatened by the significantly increased hash rate, which could be collected in the same hands. Therefore, they decided to change the algorithm. This is how the Equihash-xxx modifications appeared, where xxx is the numeric modification identifier.
The appearance of modifications coincided with a significant drop in the rate of the main cryptocurrency and the entire market as a whole, which made it unprofitable to develop new ASICs for each modification.
However, the numerous coins created on these algorithms after 2016 cannot consistently successfully compete with the top ones in terms of profitability.
If you decide to mine these coins, you will have to constantly monitor the level of profitability and periodically switch the algorithm / coin, although the profitability can be very pleasant. Especially for old cards.
These algorithms primarily require kernel overclocking.
Energy efficiency in mining is its significance. Selecting the operating mode
Any video card can work in economy mode, but they can also be overclocked as much as possible to obtain higher speeds. Although this significantly degrades energy efficiency.
Not all farms work confidently with all algorithms, so every serious miner needs to independently check the operation of their cards at the minimum and maximum, and then choose the optimal mode for them, based on the operating temperature and the cost of electricity.
The results obtained can be saved in the form of ready-made links to the 2Cryptocalc calculator, which will allow you to estimate in just a few clicks how much your chosen algorithm is better or worse than analogues in terms of the current mining profitability for your farms.
Let’s give an example of a small and very capricious farm that can confidently mine only four algorithms under Linux.
Cold mode link (430W consumption)
Hot mode link (620W consumption).
Below are the linked results for late September and late October 2020.
In addition to a noticeable one and a half-fold drawdown in profitability, it becomes obvious that it made sense to overclock this rig in September when the cost of an outlet did not exceed 3.6 rubles per kWh, and in October at a cost of no more than 2.4 rubles per kWh. At the same time, when working at the night tariff for the population – 0.8-1.8 rubles per 1 kWh, depending on the region of the Russian Federation – it was possible to work rationally in the maximum mode.
The cost of electricity and the current profitability are key factors in choosing the algorithm and operating mode of the farm.
Maximum energy efficiency and its importance when harvesting new farms
When assembling a new farm, you must clearly understand that within one series, the maximum energy efficiency of the cards is approximately the same, while most often the cards in the middle price segment of $ 300-400 get the maximum power for the price.
However, collecting and monitoring a large number of low-power farms is quite difficult and labor-intensive, and therefore the optimal choice of miners were and are cards in the price range of $ 500-700.
As a rule, all miners sooner or later come to an estimate of the maximum speed per 1 kW of farm power: someone because of the price of electricity, and someone because of reaching the limits of shareware sockets.
The maximum achievable speeds minus the devfee per 1 kW of power from the wall on a platinum PSU for the 10xx, 20xx and 30xx series are as follows.
It is easy to see that the marginal basic income from one kilowatt of energy in the 20xx series is almost everywhere 1.5 times higher than that of the previous one. This is primarily due to the improved memory access algorithm. For nuclear algorithms, the effect is not so impressive – the gain is about 33 percent.
Features of the new series of video cards. AMD 5xxx-6xxx vs Nvidia 20xx-30xx
Tight competition between AMD and Nvidia brands leads to constant updates of the lines in order to seriously outperform the competitor.
In 2017, in contrast to the successful 4xx-5xx series cards from AMD, Nvidia introduced the very successful 10xx series, which has been steadily increasing Nvidia’s market share for several years in a row.
Only by 2020, AMD were able to present the 57xx line, which seriously competes not only with the 10xx, but also the not very successful 20xx series, which Nvidia had to hastily update to the SUPER sub-series.
The announcements of the Nvidia 30xx and AMD 60xx series were striking with a sharp increase in the number of stream processors, but the first implementations of the 30xx cards did not show a multiple increase in the power of nuclear algorithms. At the same time, for memory-oriented algorithms – especially for ETH – the results are impressive. Despite the major revision of the architecture, a close approach to the limiting boundaries of modern technologies is quite obvious, which excludes further multiple development of capacities at an acceptable cost of video cards.
It remains to be hoped that in the next series there will be new technologies for working with the memory of video cards, which will help to fully unleash the potential of new and very powerful cores. An example is map 3090 and its subsequent updates).
Despite the belated start in the mining industry, the products of the Nvidia corporation confidently occupy their due share in the mining market. As of September 2020 and data from open statistics, green cards already account for 39 percent of all video cards, although in 2018 there were less than 25 percent. Of these, more than 90 percent are still cards of the 10xx family, that is, new 20xx cards are just entering the battle, and the possible massive transition of miners to the next 30xx series is on the way.
The versatility in the choice of algorithms leads to a much smaller yield gap between the top-end algorithms and alternative ones for Nvidia cards according to 2Cryptocalc, and the choice of effective algorithms itself is much wider.
We hope that a small starting educational program from this article will help novice miners get maximum profit without damaging or overheating their video cards.
The end of 2020 and the beginning of 2021 promises to be very productive with new map models. With the advent of each new card or miner programs – and maybe even new algorithms – the balance of power can change significantly, so the article will be actively supplemented as new data becomes available.