2020 bitcoin mining power consumption comparable to this country-coins, dig-IT information
United States Vice technology website published an article this week saying, due to various reasons, bitcoin network energy consumption is growing. In the most optimistic case, by 2020, the bitcoin network power consumption will reach Denmark level throughout the country . At present, we should take measures to address the problem.
Following is the full text of the article:
Kelisituofu·maermo (Christopher Malmo) recently wrote in the Vice, bitcoin development in environmental protection . This is because coins takes incredible energy. According to the calculation of Malm?, a bitcoin power Exchange reached 1.6 United States family day electricity consumption, and this number continues to grow. He pointed out: "the coming decades if treats bitcoins as major currencies, will result in the unnecessary power consumption growth, increasing human-caused climate change. ”
I also not too happy about it. According to my calculations, If the currency continues like a recent expansion of the network, the power consumption from the current output capacity growth to 2020 at a small power plant Denmark power consumption of the entire country.
So, what determines the energy of the bitcoin network?
From basic, bitcoin transactions by the decentralized system for the validation and processing. These systems typically use special hardware hashes, according to complex mathematical algorithms, to get the new currency and transaction fees.
The coins "mining machines" network to ensure security of the system, but also consumes a lot of energy. According to my calculations, present energy consumption of about 350 megawatts, the equivalent of 280,000 United States domestic electricity demand.
Mining machine efficiency
First concern is currently being used by mineral properties ("the hash is" (Hashrate), or hash, measured by the number of operations per second), as well as hardware energy efficiency (measured by hash you burn more calories at a time).
Current currency deposit machine network computing power more than 800000 trillion times per second hashing operation. More new bitcoins mine weighs more than 12 kilograms, and 1 billion times per second equipment weighing 15 grams of hash, therefore the current coins total weighing more than 10,000 metric tons of ore. These materials enough to build an Eiffel Tower.
Bitcoin development earlier, notebook and desktop computers can also be used for mining. However due to currency design mechanism, harder to dig, mine machine begin the upgrade to the video card, followed by more complex hardware. The current "the ultimate" device is the use of ASIC (application specific integrated circuit) technology to create the special coins dug chip. The emergence of this type of ASIC chips has been 3 years, which makes the device smaller, dig is more efficient.
We can see this type of ASIC mine machine efficiency improvement. Excluding ore was never formally released to the user's machine, as well as inefficiencies in the early stages of the mining machine, I listed 53 different ore coins machine, their efficiency trend is drawn in accordance with the original release date. As shown in the following figure.
Then, I would exclude USB mining machine. Because despite the high efficiency, but compared to other types of each hash of mineral prices are very high, so this type of ore proportion in the bitcoin network will not be too great.
Based on the remaining 46 mining machines to analyze trends, we can draw an optimistic and a pessimistic assumption.
Pessimistic assumptions based on ASIC miner's average. In this hypothesis, although mine function efficiency increasing, but each hash consumption of electricity is to be higher than the optimistic assumption that. The optimistic assumptions that are based on energy-efficient equipment on the market, you can see every hash is an exponential decline in electricity demand, which will lead to better long-term energy efficiency.
Although these can help us understand the future trends of the curve, but these data are based on the latest mining hardware and therefore does not represent the current energy efficiency levels in the bitcoin network. There are still a lot of older hardware in the current network. In order to determine the actual energy efficiency in the current network, I assume that hash is bitcoin network growth from newly installed bitcoin ore per month, and the year 3-5 weighted average to obtain optimistic and pessimistic assumption that overall efficiency of existing bitcoin network feeder. Below is each hashed using new hardware and the actual hardware power consumption trends over time.
Obviously, the chip vendors and hardware manufacturers with the technical progress bitcoin feeder will ensure that the future energy efficiency more than the current 3 times.
However, ore coins machine efficiency is only part of the problem. Another part of the problem is that the coins will increase in the use of it? Bitcoin Network General hash function efficiency of work force growth will offset mine better? Total energy consumption will continue to rise?
Popular coins
In order to find the answers to the questions, I check out the blockchain.info Web site provides the monthly hash calculating data. Since its January 2013 first ASIC mining machine used, hash is bitcoin network monthly growth rate of a staggering 37%. If you still maintain this growth, by the end of 2016, bitcoin network required more power than the current global capacity.
This is not consistent with the actual situation. Bitcoin sharp rise in price in 2013, and at the beginning of the end of 2013 and 2014 twice hit $ 1000. The price rise is due to abnormal speculation, so hash is based on this data to predict future growth is not accurate. We need more realistic expectations.
In the following diagram, the hash is the rapid growth in the ASIC the early popularity of the mine, while the recent growth decline.
Current 800000 trillion times per second based on the hash, I assume that the two growth rates. The first hypothesis is relatively optimistic, that is, network performance analysis within the 12-month average rate, and find out what the lowest growth since its ASIC Miner was born. This months 5%.
The alternative hypothesis is relatively pessimistic (from an environmental perspective, rather than a network security point of view), is more than 3 months before and after 12 months is also included. This months 12%.
It is clear that this rate uncertainty is large, and is always associated with currency prices: as long as the profits, mining machine computing power will continue to grow. However, the currency volatility is also very large. Therefore, a period of relatively reasonable average growth rate as a reference.
Block gains
May affect the bitcoin network hash is the growth of the other two factors were currency block size, and currency block this summer is expected to be reduced to half. Discussion continues on the block size, the conclusion is still unknown. While the blocks drop in income is due to the design of coins caused. For this reason, bitcoin transactions fees will gradually become the main motivation to dig. This can have a severe impact on energy consumption.
However, block declined to hash work force as a whole long-term impact remains unclear. The view was expressed that block drop in income will affect the bitcoin miners dig enthusiasm. Because power consumption is likely to exceed mining revenue, lead mining became a loss-making activity, this can even lead to mining activities declined substantially.
Meanwhile, the view was expressed that block drop in income will lead to scarcity of coins, currency prices then rose rapidly. This may cause the bitcoin network hash is rising sharply again. It is difficult to judge which is more likely to become a reality. Optimistic estimates are reduced by half under the block returns after 6 months, the hash is force will stop growing and stable, and pessimistic estimate is a hash based on historical levels of force will continue to grow.
Two possible scenarios
Based on the energy consumption of the bitcoin network of optimistic and pessimistic assumptions, we can draw two possible scenarios of the future development of the bitcoin network. By January 2020 or so, what kind of impact this will have on the environment? The following tables summarize the main hypotheses, calculated under various scenarios of the bitcoin network energy consumption.
Results show that under the optimistic assumption that the bitcoin network power consumption growth is not too alarming, just rise from the current 350 megawatts to about 417 MW, is within a small range of power output of the power station. If the situation got out of control, then the bitcoin network power consumption will reach 14,000 MW by 2020, equivalent to Denmark this small country's total energy consumption.
Is not serious analysis, and they also have a great deal of uncertainty. But the conclusion is important: If the currency deposit machine network continued to expand in the past, then the efficiency of mining equipment is likely to be offset, bitcoin network's total energy consumption will grow slightly, even growth.
Even on the most optimistic assumptions, turned up a currency by 2020 when the cost of energy will also be up to 5,500-kilowatt, equivalent to the United States about half of the energy consumption of households for a year. Even if we assume that only half of its electricity from fossil fuels, turned up a bit so each currency will remain more than 4,000 kilograms of carbon dioxide emissions. Due to the influence of the world so dramatically, you may query, coins are also known as virtual currency.
I still support the decentralized deal, but rethink the way they run this type of system will help, particularly for the environment. Problem solving is one of the programmes, in the system in the future, all the miners will be based on the amount of CPU processing power to pay off, but the actual hash can only be determined by thousands of randomly selected, changing CPU. We can reduce the arms race, and old mining machine with any other meaningful use.
2020年比特币挖矿耗电量将堪比这个国家 - 比特币,挖矿 - IT资讯
美国科技网站Vice本周刊文称，由于多方面原因，比特币网络消耗的能源正日益增长。在最不乐观的情况下，到2020年，比特币网络的耗电量将达到丹麦整个国家的水平。目前我们应当采取措施解决这一问题。
以下为文章全文：
克里斯托弗·马尔默（Christopher Malmo）近期在Vice上撰文指出，比特币的发展不利于环保。这是由于，比特币会消耗令人难以置信的能源。根据马尔默的计算，一次比特币交易所需的电力就达到1.6户美国家庭一天的电力消耗，而这一数字仍在继续增长。他指出：“未来数十年，如果将比特币当作主要货币，那么将导致毫无必要的电力消耗增长，加剧人为造成的气候改变。”
我的研究同样不太乐观。根据我的计算，如果比特币网络像近期一样持续扩张，那么耗电量将从当前一座小型发电厂的输出能力增长至2020年丹麦整个国家的耗电量。
那么，是什么决定了比特币网络的能耗？
从基础来看，比特币交易由去中心化的系统来验证及处理。这些系统通常采用专用硬件去进行哈希运算，按照复杂的数学算法求解，以获得新的比特币以及交易费。
这一比特币“矿机”网络确保了系统安全，但同时也消耗了大量能源。根据我的计算，目前这一能耗约为350兆瓦，相当于28万户美国家庭的电力需求。
矿机的能效
首先需要关注的是当前正在使用的矿机性能（以“哈希算力”（Hashrate），即每秒哈希运算次数来衡量），以及硬件能效（以每次哈希运算消耗的热量来衡量）。
当前比特币矿机网络的计算能力超过每秒800千万亿次哈希运算。目前较新的比特币矿机重量超过12千克，而完成每秒10亿次哈希运算的设备重量为15克，因此当前的比特币矿机总重超过1万公吨。这些材料足以建造另一座埃菲尔铁塔。
在比特币发展早期，笔记本和台式机也可以用于挖矿。然而由于比特币的设计机制，挖矿难度越来越大，矿机开始升级至显卡，以及随后更复杂的硬件。当前的“终极”设备是采用ASIC（专用集成电路）技术制造的专用比特币挖矿芯片。这类ASIC芯片的出现已有3年时间，这使得设备尺寸更小，挖矿效率更高。
我们可以看看这类ASIC矿机效率的提升。排除从未正式发布给用户的矿机，以及在早期阶段效率低下的矿机，我列出了53款不同的比特币矿机，按照最初发布日期绘制了它们的效率趋势。如下图所示。
随后，我又将USB矿机排除在外。因为尽管效率很高，但相对于其他类型矿机的每哈希运算价格也很高，因此这类矿机在比特币网络中所占比例不会太大。
基于剩下的46款矿机去分析未来趋势，我们可以得出一个乐观的假设和一个悲观的假设。
悲观假设基于所有ASIC矿机的平均情况。在这种假设中，尽管矿机能效在不断提升，但每哈希运算消耗的电力仍要比乐观假设中更高。而乐观假设基于市面上能效最高的设备，可以看到每哈希运算的电力需求正呈指数式下降，这将带来更好的长期能效。
尽管这些曲线能帮我们了解未来的趋势，但这些数据基于最新挖矿硬件，因此并不代表当前比特币网络中的能效水平。当前网络中仍存在大量较老的硬件。为了确定当前网络中的实际能效，我假设比特币网络哈希算力的增长完全来自每月新安装的比特币矿机，并对3到5年的情况进行加权平均，以求得乐观和悲观假设下网络现存比特币矿机的总体效率。下图是每哈希运算采用新硬件和实际硬件所消耗电力的长期趋势。
很明显，芯片厂商和硬件厂商的技术进步将确保未来比特币矿机的能效超过当前的3倍。
不过，比特币矿机的能效只是问题的一部分。问题的另一部分在于，比特币的使用量是否会出现增长？比特币网络总体哈希算力的增长是否会抵消矿机能效的提升？总能耗是否仍将继续上升？
受欢迎的比特币
为了找到问题的答案，我查阅了blockchain.info网站提供的月度哈希算力数据。自2013年1月第一台ASIC矿机投入使用以来，比特币网络哈希算力的月增长率就达到了惊人的37%。如果目前仍保持这样的增长，那么到2016年底，比特币网络所需的电力将超过当前全球的发电量。
这并不符合实际状况。比特币价格2013年曾大幅上涨，并于2013年底和2014年初两次触及1000美元。这一价格上涨是由于异常的炒作，因此基于这一数据去预测未来的哈希算力增长并不准确。我们需要更符合实际的预期。
在以下这张图表中，哈希算力的高速增长出现在ASIC矿机的普及之初，而近期的增速则出现下降。
在当前每秒800千万亿次哈希运算的基础上，我假设了两种增长率。第一种假设相对乐观，即分析12个月内平均的网络性能增速，并找出自ASIC矿机诞生以来的最低增速。这相当于月增速5%。
另一种假设相对悲观（从环保角度，而非网络安全角度来看），即将以上12个月的前后3个月也包括在内。这相当于月增速12%。
很明显，这一增速的不确定性很大，并且总是与比特币价格相关：只要能带来利润，矿机的计算能力就会一直增长。然而，比特币的价格波动性也很大。因此，采用一段时期的平均增长率作为参考相对合理。
区块收益
可能影响比特币网络哈希算力增长的另两大因素分别为比特币区块的大小，以及比特币区块收益预计将于今年夏季降低至一半。关于区块大小的讨论仍在继续，结论仍然未知。而区块收益的下降则是由于比特币的设计造成的。基于这一原因，比特币交易费将逐渐成为挖矿的主要动力。这可能会对能耗造成严重影响。
不过，区块收益下降对哈希算力整体的长期影响仍然不明。有人认为，区块收益的下降将影响比特币矿工的挖矿热情。由于耗电量可能将超过挖矿收益，导致挖矿成为一项亏损的活动，这甚至会导致挖矿活动大幅减少。
与此同时，也有人认为，区块收益的下降会导致比特币的稀缺，进而造成比特币价格迅速上涨。这可能导致比特币网络哈希算力的再次大幅上升。目前还很难判断，哪一种情况更可能成为现实。乐观估计是在区块回报下降至一半的6个月之后，哈希算力将停止增长并趋于稳定，而悲观估计是哈希算力仍将继续按照历史水平增长。
两种可能的场景
基于对比特币网络能耗的乐观和悲观假设，我们可以得出未来比特币网络发展的两种可能场景。那么到2020年1月左右，这对环境将产生什么样的影响？以下表格总结了主要假设，计算了各种场景下比特币网络的能耗。
结果显示，在乐观假设下，比特币网络电力消耗的增长不会太令人震惊，只是从当前的350兆瓦上升至约417兆瓦，仍处于小型发电站的电力输出范围内。如果情况失控，那么比特币网络到2020年的电力消耗将达到14千兆瓦，相当于丹麦这种小国的总能耗。
以上并非严肃分析，而这些数据也有很大的不确定性。不过结论很重要：如果比特币矿机网络继续以以往的方式扩张，那么挖矿设备的效率提升很可能将被抵消，比特币网络的总能耗将会小幅增长，甚至大幅增长。
即使是在最乐观的假设中，到2020年挖出一个比特币耗费的能源也将高达5500千瓦时，相当于美国普通家庭一年能耗的约一半。即使我们假定，届时只会有一半电力来自化石燃料，那么每挖出一个比特币排放的二氧化碳仍将超过4000千克。由于对世界的影响如此明显，你可能会质疑，比特币是否还配称作虚拟货币。
我个人仍然支持去中心化的交易，但重新思考这类系统的运行方式将带来帮助，尤其是对环境而言。解决问题的方案之一是，在未来的系统中，所有矿工都将基于投入的CPU处理能力获得回报，但实际的哈希运算只会由数千个随机选择、不断变化的CPU去进行。我们可以减少计算能力的军备竞赛，并将旧矿机用于其他更有意义的用途。