Mining history. Part I – The Evolution of Bitcoin Mining Devices

Many respected economists recognize that the emergence of cryptocurrencies can be the starting point for a radical transformation of the entire global financial system.. But financial technology is not the only area where bitcoin has created an incentive for development and improvement.. Another area is the extraction of bitcoins, known to us as mining. Indeed, the evolution of mining devices in just a few years is amazing.. In fact, “coin-mining” In just three years, the industry has created a new class of computing machines and has gone from the long-obsolete 130 nm process technology to the cutting edge of modern electronics. In the first part of the article, we will talk about the development of mining technologies from the launch of the Bitcoin system to the start of mass production of specialized devices. Despite the fact that more than half of the period of existence of Bitcoin, mining took place on general-purpose computers, we will not dwell on a detailed description of this process.. It is unlikely that you will be interested in reviews of the most common processors and video cards. But the story of the rapid development of ASIC miners, fierce competition, the rapid enrichment and ruin of manufacturers, the development of enthusiasts and squeezing all the possibilities out of new and old hardware is actually fascinating, even for those who once took part in it and now, in the noise of the fans or the murmur of the cooler, continues to mine new coins. Start. Mining on processors It is easy to guess who became the very first miner in the history of Bitcoin. Of course, it was the creator (or creators) of the cryptocurrency – Satoshi (Satoshi) Nakamoto. It is widely believed in the community that Satoshi has been mining alone for a long time and has over a million BTC in his wallets. However, it is very difficult to substantiate this opinion with facts.. The only evidence they give is that all these coins have not yet been set in motion.. But there can be many reasons for this. We will operate only with known facts. Between the creation of the genesis block on January 3, 2009 and the publication of the first client of the Bitcoin v0.1 network, only 15 blocks were mined – from 0 to 14. Thus, the “guaranteed” Satoshi premine is only 750 BTC. Further, other miners could already compete with him. Now these people are called “early adepts” – early adopters. Most of them are unknown to the community. The first transaction in the Bitcoin network occurred 3 days later – on January 12, 2009 between Satoshi and Hal Finney in the amount of 10 BTC. In total, 5 transactions were made from this address, which undoubtedly belongs to Satoshi, on the same day, for a total of 32 BTC. From a technical point of view, Bitcoin mining on processors is of no interest – this is a common hash calculation operation using the SHA256 algorithm, which is also performed in many other cases not related to cryptocurrency. Probably, Satoshi did not imagine how quickly the mining industry would develop.. It was mining on PC processors, the most massive chips in the world, that was supposed to make Bitcoin truly decentralized.. There is only one vulnerability here – in front of botnets, which can force computers of tens of thousands of users to mine computers for one wallet. While cryptocurrencies were fun for geeks, CPU mining was not very popular.. The first change in the complexity of Bitcoin, which showed that several hundred processors are already mining, happened almost a year after the creation of the genesis block – on December 30, 2009. The largest increase in complexity in one recalculation – 4 times – happened on July 16, 2010 in block 68544. This happened as a result of the publication about Bitcoin on July 11 on the site Slashdot, popular among geeks.. In just a few days, the number of Bitcoin users, and therefore miners, has grown many times over. Farmers Contest. Mining on video cards By the summer of 2010, the popularity and exchange rate of Bitcoin had grown so much that its mining began to bring real income, albeit a modest one.. In July, for 1 BTC they gave about 10 cents, that is, the extraction of one block brought about $5. Mining began to switch to a commercial footing, which means that competition could not but lead to a technological race. On July 18, someone ArtForz launched a mining farm on the GPU for the first time and mined the first block using parallel computing implemented in the OpenCL driver. Thus began the era of industrial mining. The transition to video cards was a huge breakthrough compared to the “classic” processor – not only did one video card count several tens of times more hashes than a processor, even then it was possible to install up to 4 video cards on one cheap motherboard, and later – up to 6 , limited to 8 GPUs (one dual-socket card counts as two GPUs). A computer even with two processors in the minimum configuration costs almost the same as a farm of several top-end video cards. Thus, video cards immediately relegated processors to the background. But for some time, programs for mining on the GPU remained unavailable for mass use.. Only in September 2010 was a CUDA-based miner for nVidia cards published, and in October for ATI Radeon based on OpenCL. On September 18, the first pool in the history of cryptocurrency was opened – Bitcoin.cz, known as Slush's pool. It still works, and its creator, Czech programmer Marek Palatinus, is still an active member of the community.. In the first days of January 2011, the pool gained a capacity of 10 Gh / s, which now seems ridiculous. But back then Bitcoin mining difficulty was barely over 10,000 – only 5 million times less than it is now! On February 9, 2011, the Bitcoin exchange rate on MtGox caught up with the dollar, and overcame the $20 bar for several months without stopping.. The extraction of one block now brought $ 1,000, and almost 150 thousand dollars were mined per day. Feeling easy money, miners around the world rushed to buy video cards. The competition was growing, the difficulty continued to rush up, reaching 1,000,000 by mid-June. However, the hack of MtGox, and then several other services, caused an outflow of Bitcoin users and a decrease in mining capacity, which continued until the fall. On August 23, 2011, the first block was mined by a decentralized pool – P2Pool. Unfortunately, due to the complex setup and instability of mining, most miners still prefer centralized services, and p2pool drags out a miserable existence, very rarely mining more than 1 block per day. October 7, 2011 was launched Litecoin – the first fork of Bitcoin on an alternative hashing algorithm. It was intended to stop the dominance of video cards and give a chance for mining on the CPU.. Ironically, after 2 years, it was LTC that became a refuge for miners on video cards fleeing the invasion of ASIC. We will talk about this in another article. For almost three years, video cards reigned supreme in mining. November 28, 2012 saw the first block reward halving from 50 BTC to 25 BTC.. But it did not stop the gradual increase in complexity.. At the beginning of 2013, despite a slight correction, it fluctuated around 3,000,000. However… it was only preparation for a new jump. Smaller, colder, more powerful. FPGAs come and go Back in 2011, some enthusiasts began to realize that graphics card farms consume too much electricity, require constant attention and additional costs.. They began looking for a solution that could cut costs. At that time, the most obvious was the use of FPGA chips – less versatile than CPUs, but more energy efficient.. It was much easier to parallelize them on a single board, and the requirements for power lines and cooling were noticeably reduced.. The end device turned out to be quite expensive, but much more compact and stable than a farm on video cards. And the gain in energy consumption was thousands of percent. And yet, video cards remained the most massive solution, cheap and affordable.. There were few people who were willing to pay several thousand dollars for a device with a very limited use.. FPGA miners did not last long and remained a niche product that did not play a significant role in mass mining. But these developments were then useful to the manufacturers of ASIC miners, since the layout of the boards, the layout of the devices and the software were very similar. Among the manufacturers of FPGA miners, two companies can be noted, which later took a much more prominent place in history: Swedish KnC Miner, founded in September 2012, with its KnC Mars with a power of 6 Gh / s American Butterfly Labs with two devices: BitForce SHA256 Single with hashrate 832 Mh / s and Mini Rig with a hashrate of 25 Gh / s. For that time, it was very powerful equipment, consuming only 20 W per 1 Gh / s – 20-30 times less than a similar farm of video cards. But it cost accordingly: they asked for $6,000 for Mars, and $15,000 for Mini Rig. The number of released devices was not published. All other manufacturers could not recoup the costs, and even more so the development of new chips. Over time they all closed.. The payback of FPGA miners raised many questions, however, those who received them before the beginning of 2013 had every chance, if not for super profits, then for a decent profit. Advent. The first ASIC miners Unlike FPGAs, which are mass-produced and used for a variety of tasks, ASIC (Application Specific Integrated Cirquit) chips can only perform one task for which they were designed. But they perform it much better than any general-purpose processor – the difference in performance of devices similar in other parameters can differ by dozens of times. There is a downside – only the design and production of a prototype ASIC chip using modern technologies costs several million dollars.. This is a long and laborious process, the positive result of which is not guaranteed by anyone.. In addition to the chip itself, the developer has to do everything – from the motherboard, the “strapping” and the cooling system to the software, and then carry out all the stages of testing. The most far-sighted manufacturers began developing specialized chips as soon as the first FPGAs went into production – that is, in the summer-autumn of 2012. And the most enterprising decided that there was no need to take all the risk on themselves – buyers can perfectly pay for it. This is how pre-orders for ASIC miners appeared. The first well-known trio of ASIC creators were two Chinese companies – ASICminer and Avalon and one American – BFL (Butterfly Labs). The creators of the first ASICs, from left to right: Sam Cole (KnC Miner), Josh Zerlan (Butterfly Labs), FriedCat (ASICminer) [allegedly] and Yifu Guo (Avalon) In the photo you can see four people – probably the most famous during the “ASIC -revolution”. It was made in Singapore in August 2013 and is truly unique – this company has never been together again.. Remember these names and faces. With these people, wholesale and retail, hundreds, if not thousands, of ASIC miner buyers dreamed of a heart-to-heart talk. But let's talk about each of the manufacturers in order. ASICminer This company with a telling name was founded on July 18, 2012 by three citizens of China. She collected investments through public forums, but not from the first people she met. ASICminer had a completely different approach to clients than other manufacturers. For a long time they did not trade at retail, but focused on large investors who received their share of the shares. On the open market, these shares began to sell only after the main investors received their piece of profit.. Representatives of the company answered absolutely all questions, including those related to deeply technical issues. The identity of the main founder and inspirer of ASICminer still remains unknown – in the community he is known only by the nickname “FriedCat” (Fried cat). Perhaps he is not Chinese, although he managed to settle down and build a business in this country. Employees of the company do not betray their boss. At the beginning of 2013, ASICminer developed a chip using the 130 nm process technology, which has long been outdated in consumer electronics.. It was chosen because of the relatively low cost – in total, $150,000 was spent on the development of a completely new ASIC. Miners of the first generation from ASICminer were implemented in the form of Blade Block Erupter boards with good characteristics for that time: a hashrate of 10 Gh / s and a power consumption of about 100 W. The energy efficiency indicator, despite the rough process technology, was twice as good as that of the most modern FPGAs. Block Erupter boards could be installed in multiples in one chassis, compatible with standard 19″ server racks. It was ASICminer who became the first company to open a special data center for mining. In the spring of 2013, devices of the first batch were installed in it, which were sold out in 1 hour at a price of $ 12,500 by the company's shareholders. Their total performance was up to 40 Tx/s, while on January 1, 2013 the hash rate of the entire Bitcoin network was about 25 Tx/s. For several months, this data center held the absolute leadership in Bitcoin mining, ahead of the largest pools. Blade Block Erupter v.2 sold in April for $7,500 on the open market. The miner turned out to be successful, so the company decided to improve ASIC and in the summer began large-scale sales of ASICminer Cube at a price of $7,000. Its characteristics: 30-37 Gh / s and 430 watts. Simultaneously with Blade Block Erupter, a portable version was released – USB Block Erupter with one chip, in a case similar to a flash drive. A key fob with a capacity of 330 Mh / s was first sold for 1 BTC and at first the demand was booming – they were bought in dozens and connected through USB hubs in whole bunches. However, the idea was unsuccessful, the miners were rapidly becoming cheaper – in a few months their price collapsed by 30-40 times. Now such a curious device can be bought on Ebay for a few dollars.. This souvenir is also suitable for studying the operation of ASIC chips and circuit design. By the end of the summer of 2013, ASICminer was forced to retreat under the onslaught of competitors – its chips were quickly outdated, and the development of new ones was slow. Shares fell. The company was able to return to the market only a year later, but soon closed completely after the mysterious disappearance of the Fried Cat with a large amount of money.. He was never found. Avalon Project Avalon sounds good, doesn't it? That's right, this company has something to be proud of. It was founded by Yifu Guo in September 2012. Although Yifu is a full-blooded Chinese, his firm's association with King Arthur's fairytale city suggests that he is well-acquainted with European culture and relied primarily on European and American buyers. The Avalon team began developing chips at the same time as competitors, but produced only limited batches at a high price.. The so-called “batches” from Avalon stirred up rumors and legends in the community.. The first ASIC miners were released on the 110 nm process technology, with a default hashrate of 60-65 Gh / s (82-88 Gh / s overclocked) and a power consumption of about 700 watts. The release of the first generation took place in three stages: The first batch of 300 devices was sold at the very start – in 2012 for $ 1300, when the idea seemed too bold, and the start of pre-sales did not cause much excitement. Buyers received miners in February-March 2013. They paid off literally in a matter of days and began to bring net profit.. The owners of the first batch of Avalon really got rich, and their example became a bait for buyers of the following batches.. But they were far less successful. The second batch, already from 600 miners, went on sale in mid-February for $1,500 (at the exchange rate in bitcoins – at that time about 50 BTC) and was sold out in a few days, despite the generally accepted skepticism about ASICs at that time.. Here, buyers were much less fortunate – the devices arrived with a long delay and had already been in use.. However, they turned out to be a good purchase.. Read a detailed review of one of them in this article. But the third batch, also of 600 pieces, greatly tarnished the reputation of the manufacturer. Firstly, the price rose almost 5 times – 75 BTC at an average rate of about $100. And secondly, the delay in deliveries turned out to be even greater – many buyers received their devices in July and even in August, when the complexity increased many times over and second-generation ASICs entered the market.. Most owners of third-party Avalons lost money and, not without reason, blamed the manufacturer for this. Miners from Avalon work stably, but unfortunately the manufacturer uses an unsuccessful TP-Link controller, which is distinguished by its underdevelopment and constant “freezes”. This shortcoming accompanied Avalon devices for more than one year.. However, the company has not slowed down and still remains in the ranks after the transformation into Canaan Creative, although it has lost the advantages accumulated at the beginning – now Avalon is just one of many manufacturers. BFL – Butterfly Labs BFL, aka “Butterfly” and “Grandma Labs” is one of the most controversial in the history of mining, its name has become a household name to describe an unreliable supplier that does not fulfill its obligations. It undeniably and by a significant margin has secured its place as the best manufacturer of pre-orders for ASIC miners.. Only getting ready-made devices turned out to be much more difficult. Butterfly Labs became the first manufacturer to start ASIC pre-sales. Confidence in success was inspired by previous developments in the field of FPGA miners, which were delivered on time and worked well. In June 2012, the company offered customers the following lineup: $149 Jalapeno with 4.5 Gh/s SC Single for $1299 and 60 Gh/s SC MiniRig for $30,000 and 1500 Gh/s one day, pre-orders worth $250,000 were redeemed, and in the future, the interest of “investors” did not fade away for a long time. The chips were promised with 65nm process technology, energy efficiency of 0.9W at 1Gh/s and QFN package. But when we got down to business, it turned out that creating an advanced ASIC is much more difficult than collecting pre-orders.. Shipment dates have been postponed many times, energy efficiency has changed to 5.5 W at 1 Gh / s, the package has changed from QFN to BGA, and the power of SC MiniRigs has decreased to 500 Gh / s – instead of one “box” customers were promised 3! The Internet was full of angry reviews, because the company did not always report the reasons for the next failure to meet deadlines, and very few lucky people managed to receive compensation.. The phrase “just two weeks” (two more weeks), which BFL PR director Josh Zerlan tried to reassure customers, has become winged among the miners. The first Jalapeno devices appeared in reviews at the end of April and began to arrive to customers only in May 2013, and bulk shipments went in June – July. In August, the growth of complexity reduced the profitability of BFL devices to the payback line. And those who stood at the end of the line received orders in the last months of 2013, when the miners from BFL, compared to competitors, were only suitable as electric heaters. In August 2013, the company announced a unique 28 nm ASIC – Monarch, made in the format of a video card, inserted into a regular PCI-e slot of a personal computer, but with an external power supply. The second delivery option was with a USB connection.. “Monarchs” were sold at the following rates: 300 Gh/s for $1497 and 600 Gh/s for $2196. Unlucky buyers who were still waiting for Single and MiniRig ordered a year ago were offered to convert their orders to “Monarchs” with an additional payment and wait another N * “just two weeks”. There are few such people. In March 2014, deliveries of the Monarchs did begin, but in very limited quantities. Butterfly Labs no longer had the funds for mass production. The company was filed with several claims from buyers for delays and refusal to return, and at the end of 2014 it was closed for several months by the FTC – the US Federal Trade Commission. But in January 2015, it reopened and even issued a certain amount of compensation. Further prospects for Butterfly Labs did not inspire optimism – the trust of the community was completely lost, except for minor improvements to the Monarch, there were no new developments. The BitSafe hardware wallet has not left the testing stage. Soon the “grandmother” of cryptocurrency mining quietly died. But the history of mining does not end there. To be continued…