Perhaps on the Internet you have met heated discussions of the upcoming update on the Ethereum blockchain, which relate to the new ProgPoW algorithm. This topic has remained quite topical for over a year, because the new software will fundamentally reverse the mining process in the mentioned network.
In the beginning, almost 100% of asset holders were categorically opposed to innovation, but over a long time passions calmed down a bit, and the audit finally put all the points on the i – algorithm to be. Let’s find out in more detail what kind of technology it is, why it is needed and why it causes such a stormy reaction.
The update suggests making changes to the current Ethereum Ethash mining algorithm. The key difference of the new development is the dynamic change of the cryptographic task every 12.5 minutes, which will deprive the ASIC miner systems of their advantages and open new horizons for the GPU.
If, or rather when, the proposed code changes are accepted, this is likely to reduce the number of ASICs in the Ethereum network. Code change proponents claim that ASICs are one of the most serious threats to decentralizing the Ethereum network.
And this view is popular despite the fact that ASIC is a relatively new phenomenon on the Ethereum network. Bitmain, ASIC’s largest hardware provider, shipped its first batch of ASICs destined for the Ethereum network in July 2018. Regardless, the ASICs developed for Ethereum represent only a small fraction of the products offered by ASIC hardware vendors, so the impact on these vendors’ business models will be negligible.
However, GPU vendors such as Nvidia and AMD can expect a significant increase in sales. In fact, it is unclear what proportion of mining these plants currently have, but the new algorithm will definitely make it even smaller. This, combined with the planned transition to PoS, creates many problems for large mining farms.
Why do you need ProgPoW
The main reason for the development of the ProgPoW algorithm is the influence that ASIC has on the cryptocurrency industry, which jeopardizes the main cryptocurrency principle – decentralization. But at the same time, the current Ethash algorithm, which is expected to be replaced by ProgPoW, is considered the most ASIC-resistant algorithm.
Therefore, it is quite difficult to understand the need for another algorithm that aims at the same result. On April 3, 2018, Bitmain officially announced the new powerful ASIC model developed for ETH mining, which inspired other manufacturers to do the same. This led to unpleasant consequences for the Ethereum community, mainly related to security and integrity issues. First of all, owners of GPU farms began to suffer losses and lost interest in servicing the network. In 2018, the income of video card manufacturers has fallen and has since been steadily declining.
Overall, GPU shipments fell by 3.3 percent annually. Susquehanna, a global technology company, said its monthly earnings from mining ETH using GPUs fell from $ 150 in the summer of 2017 to zero in November. The profitability of mining using GPUs fell as the price of ETH fell. Moreover, the main computing power was at the disposal of certain groups of people, which jeopardized decentralization and network security.
The reason is the dominance in the market of such giants as the same Bitmain, which created its monopoly by issuing new ASICs for top coins. Since then, the ecosystem has become increasingly centralized and vulnerable to 51% attack.
The last straw was when the Bitmain developers announced the release of ‘the most powerful and efficient EtHash ASIC in the world’ for Ethereum and Ethereum Classic mining. Since then, the Ethereum Foundation has been working to solve this problem, and as a result, the ProgPoW proposal has appeared.
How effective is code change
Improving the competitiveness of GPU cards and reducing centralization will be possible due to the inconsistent mining condition and the use of all resources and capabilities of video cards. The problem is that ASIC settings are designed for one specific task: calculating blocks for cryptocurrency mining.
Meanwhile, video cards are more flexible and can perform a number of tasks from mining to image transfer. The randomness of the set puzzles will not allow special-purpose integrated circuits to work in the usual uninterrupted mode based on the template. In simpler terms, the new algorithm requires flexible mining.
The developers plan to achieve such results through five innovations, in which the algorithm will be based on the power of computing resources, the speed of processing traffic and the amount of memory.
- Changes in the variable bit length hashing algorithm. The SHA-3 hash function size will decrease from a 64-bit f1600 to a 32-bit f800. The latter has been optimized for 32-bit platforms to better match the GPU architecture.
- Increased mixing matrix coefficient. This is the number of processor registers in which intermediate values can be stored. They are significantly faster than RAM, because they are located inside the processor. In integrated circuits, the chips are small, and the number of registers is lagging behind GPUs in performance. As a result, ASICs are forced to use RAM to work with this function, which is several times slower.
- Adding a random sequence of calculations to the main loop. The integrated circuit makes it impossible to create an ASIC device with a fixed pipeline, which could increase the speed of work.
- Adding read operations from an intermediate buffer with low latency and support for arbitrary addresses. This forces the ASIC to follow the rules of the GPU memory hierarchy and limits their capabilities and performance.
- Random access dynamic memory increased from 128 to 256 bytes. Video cards are available in large volumes, and specialized ASICs are not able to optimize the memory controller to improve performance.
Users are still arguing about the feasibility of introducing a new algorithm. On the one hand, ASICs really have a tendency to centralize power, but on the other hand, these installations, and especially mining farms, are incredibly expensive, which to a large extent guarantees the interest of their owners in the operation of the blockchain.
A network of many small miners, who can now easily afford the necessary equipment, clearly seems more decentralized. But, firstly, there is a high probability of the appearance of GPU farms, and secondly, this type of miners will not be so devoted to the Ethereum network, since initially it did not incur any costs for mining, and therefore does not risk anything.
Therefore, even leading experts do not undertake to make any accurate forecasts, but it is highly likely that the ProgPoW algorithm will become only a transition point on the way to PoS.