OverviewBitmain recently launched a new Ethereum miner, widely believed to be an ASIC, which is expected to ship by the end of July 2018. However, many in the Ethereum community are opposed to ASICs and prefer GPU mining, as the companies that produce GPUs are primarily focused on the computer gaming industry rather than the crypto industry, which should mean that hardware distribution will be more widespread and fair, thereby improving decentralization. Therefore, the risk for Bitmain could be that the Ethereum community may decide to change the PoW algorithm with a hard fork, which could devalue Bitmain's miners and cause its R&D investment to go to waste. In this report, we speculate that Bitmain may have taken a head start compared to the Ethereum community. Bitmain may have learned from Monero, which recently underwent a PoW change that caused Bitmain's ASIC chips to depreciate significantly. Developing a specific chip requires a considerable investment, so we believe that Bitmain may have taken some countermeasures to avoid repeating the same mistakes. Bitmain may have designed a new type of mining chip that is less efficient as an ASIC but will not be affected by the PoW change. This would make Ethereum's hard fork for the PoW change meaningless. Monero recently made changes to resist ASIC PoWIn early April 2018, the Monero community decided to change the PoW algorithm through a hard fork in an attempt to render ASICs useless and make Monero more GPU-friendly. Due to the sharp increase in hash rate, as shown in Figures 1 and 2, the Monero community believed that ASIC manufacturers had secretly developed Monero ASICs and were mining Monero tokens. As shown in Figure 2, the 90-day rolling hash rate growth rate in early 2018 reached about 300% (calculated based on a 7-day rolling average). This growth rate is extraordinary even considering the sharp rise in the value of Monero. After the Monero programmers announced the plan for the hard fork, Bitmain began selling Monero ASICs on their website, indicating that they may have been mining in secret. As shown in Figure 1, the Monero hash rate dropped significantly after the PoW change. After the hard fork, Monero split into two chains, and the original regular coin is called the Monero original chain (XMO). Although this coin is worth less than Monero, it has a higher hash rate because there are no other coins to mine besides the Monero original chain. Although this fork does not have replay attack protection, Monero has added ring signature restrictions, so Monero and the Monero original chain can be split by a small number of (less than 7) ring signature transactions on the Monero original chain. Figure 1 - Monero hash rate compared to price Source: Coinmarketcap, BitMEX Research Chart 2 – Monero’s hash rate compared to price – 90-day rolling percentage of the 7-day moving average Source: Coinmarketcap, BitMEX Research Note: The rolling average of hashrate within 7 days after the PoW hard fork excludes data before the hard fork. Bitmain's new Ethereum mining machineAs we mentioned above, Bitmain recently launched a new Ethereum miner that is expected to ship by the end of July 2018. Bitmain may be concerned given the history of Monero and the fact that many in the Ethereum community (including those who mine Ethereum at home with GPUs) may be unhappy with Bitmain's new product. One downside of the new miner is that it may encourage centralization of miners, but beyond that, the existing miners' existing financial interests in holding GPUs may also create hostility towards the product. Bitmain's management is not stupid, so we think the company is likely to act cautiously and may have taken steps to mitigate some of the risks. Figure 3 - Bitmain’s new Ethereum mining machine: Antminer E3 Mining machine specifications:
Source: Bitmain The specifications of this product are as above. As shown in the table below, this new Ethereum miner is less efficient than one would expect from an ASIC miner when compared to the efficiency gains of other coin ASICs. For example, Bitcoin ASICs are about 521 times more efficient than FPGAs, while Monero ASICs are about 88 times more efficient than GPUs. In comparison, the new Ethereum miner is only about 1.4 times more efficient than a GPU. This may indicate that the new Ethereum miner is not an ASIC at all, but simply a new device that is more efficient than existing GPU miners. However, while the table below is a rough calculation that ignores many key variables and factors, such as the memory-intensive nature of the Ethereum mining algorithm, the numbers can still illustrate the point, although the calculation results are not absolutely precise: Figure 4 - Miner efficiency calculation (approximate value)
Source: BitMEX Research, Bitmain Note: Data are approximate Mining chip types and vector processors (VP)As shown in Figure 5 below, when Bitcoin was launched in 2009, CPUs were generally used for mining. However, the system structure of GPUs and FPGAs is more efficient when processing repeated hashing operations. Therefore, the network first moved to GPUs and then to FPGAs. In 2013, ASICs for specific hash functions appeared. Compared to CPUs, GPUs, and FPGAs, ASICs are more efficient when running specific hashes, but the efficiency of ASICs on other hashes is very low or actually completely useless. Figure 5 - Timeline of the development of cryptocurrency chip types Source: BitMEX Research Note: Whether the vector processor (VP) can be put into use by the end of 2018 remains to be seen. Bitmain may have developed a new type of chip – a vector processor. The architecture of this chip can be used for PoW hash functions in general, but not for a specific one. These chips may be more efficient than GPUs and FPGAs, but less efficient than ASICs. The advantage over ASICs is that they may be in some ways less vulnerable to PoW updates. The new Ethereum miners are based on this type of chip, even if this is just our guess. Figure 6 - Evolution of Cryptocurrency Chip Types
The higher the rate Bitmain's new Ethereum miners are likely to be tailor-made for Ethash, as the components inside the miners such as circuits, power control devices, memory and control modules can all be calibrated specifically for Ethereum mining. However, the chips themselves, which are by far the most in need of investment, are likely to be more general and not designed specifically for Ethereum. Therefore, if Ethereum undergoes a PoW change, the chips can be assembled into new devices when they leave the manufacturer, or it is even possible to use chips from old devices in new Ethereum miners. Although this is still just speculation on our part. Artificial Intelligence (AI) TechnologyOn April 19, 2018, at TSMC’s new quarterly earnings report, Co-CEO Mark Lie said: [Bitmain] has made many contributions to blockchain technology, such as AI. They have done a good job. We expect them to slowly move into the AI field. Source: Q1 2018 earnings call “AI” is a broad term. While it’s yet to be confirmed, the new vector processor chip could be what TSMC refers to as “artificial intelligence technology.” Because such chips can switch between hashing algorithms, we can assume that they fall into the AI category to a certain extent. Whether vector processor chips are programmable, like modern GPUs, or generally more efficient than GPUs remains to be seen. If such a chip does appear, this advanced technical capability could be seen as a major achievement for Bitmain. This technology could be more expensive to develop and more specialized than ASIC technology, which could make mining centralization more problematic. Ethereum hash rate growth – evidence that new chips are still not deployedDespite this, we have not seen any interest in deploying new chips on the Ethereum network. As shown in Figures 7 and 8 below, Ethereum’s hashrate (in a broad sense) continues to follow normal trends despite price fluctuations. Figure 7 - Ethereum hash rate compared to price and Nvidia GPU sales Source: Bloomberg, Etherscan.io, Coinmarketcap, Nvidia, BitMEX Research Chart 8 – Ethereum Hashrate vs. Price – Rolling 90-Day Percentage Growth of 7-Day Moving Average Source: Bloomberg, Etherscan.io, Coinmarketcap, BitMEX Research in conclusionWhen discussing whether Bitmain’s new Ethereum miners are ASICs and whether the new chips might be immune to PoW changes, Vitalik Buterin told us:
Despite what we have said above, much of this article is still based on our speculation. However, even if we are wrong about this particular chip, we still think that in the future Bitmain or another company may develop a general purpose hashing chip that will be more efficient than a GPU for any hashing algorithm. At this point, the era of fighting ASIC PoW changes may be over, and the crypto community must choose between two bad apples:
Unless the PoS system is proven to be strong enough.
This article is provided by the BitMEX research team. For more research reports on the cryptocurrency industry, please visit http://cn.research.bitmex.com Author: BitMEX Research Team Weibo: BitMEX WeChat Official Account: BitMEX Original article: https://blog.bitmex.com/zh_cn-new-ethereum-miner-could-be-a-game-changer/ |
<<: Latest firmware for A8 &A8+ &S11
>>: UUpool mine pool launches BTM mining
Author | Hashipi Analysis Team...
At this very moment, 32 central banks and 12 fina...
On June 5, a Bloomberg news triggered heated disc...
There are such people in our lives. Their fortunes...
The zygomatic bone is one of the important bones ...
It has been 20 days since the plunge on August 2....
Career line in palmistry, career line in palmistr...
In physiognomy, people with moles on their chest ...
One of the factors that influence destiny is the ...
It seems that everyone wants to live a long life,...
1. Black tear troughs are definitely a sign of ki...
Both boys and girls can have moles, so both men a...
In interpersonal communication, people don't ...
Source: Ti Capital, original title: Taihe Observa...
The lifeline is the line that runs from the botto...