The four most famous international exchanges:

Binance INTL	OKX INTL	Gate.io INTL	Huobi INTL
China Line　APP DL	China Line　APP DL	China Line　APP DL	China Line　APP DL

Note: The above exchange logo is the official website registration link, and the text is the APP download link.

As the official operation of Ethereum approaches, more and more friends are asking how to mine Ethereum. This article synthesizes the information we know so far and hopes to be of some help to everyone. Since the design and improvement of the mining algorithm is still in progress, the following information is for reference only.

 Design goals

Resistance to miners (ASIC Resistance). The mining advantage produced by using specially optimized chips should be as small as possible, so small that even mining with an ordinary CPU can generate profits.

Light client verifiable. Light clients should have the ability to verify the authenticity of every block. The goal is to run the verification algorithm implemented in C on an ordinary desktop computer. The verification time is less than 0.01 seconds, and less than 0.1 seconds using Python or Javascript, and the memory usage does not exceed 1MB.

 Mining Algorithm

The mining algorithm of Ethereum is called Ethash, also known as Dashimoto (Dagger-
Hashimoto) is a variant of the Hashimoto algorithm combined with Dagger. Its characteristic is that mining efficiency has basically nothing to do with the CPU, but is positively related to the memory size and memory bandwidth. The requirements for memory size and bandwidth mean that mining chips deployed on a large scale through shared memory cannot have linear or super-linear mining efficiency.
linear) growth.

 Next let’s take a look at Dashimoto’s basic process:

For each block, a seed is first calculated. The calculation of the seed only relies on the current block information, such as block number and block headers.

Use the seed to generate a 32MB pseudo-random data set, called cache. Light clients need to save cache.

Based on the cache, a 1GB data set is generated, called the
DAG. Each element in this data set only relies on certain elements in the cache. In other words, as long as there is a cache, the element at the specified position in the DAG can be quickly calculated. A complete minable client requires a saved DAG.

Mining can be summarized as the process of randomly selecting elements from a DAG and then hashing them. The verification process is the same, except that instead of selecting elements from the DAG, the element at the specified position is calculated based on the cache, and then it is verified that the hash result of this element set is less than a certain value. Since the cache is very small,
And the DAG elements at the specified position are easy to calculate, so the verification process only requires ordinary CPU and ordinary memory to complete.

The cache and DAG are updated every cycle, and the length of a cycle is 1000 blocks. In other words, the cache and DAG generated by these 1000 blocks are exactly the same, so the main work of mining is to read data from the DAG rather than update the cache and DAG. The size of the DAG grows linearly over time, starting at 1GB and increasing by approximately 7GB per year
– So by December 2015 it was about 8GB, by December 2016 it was about 15GB.

Note: The numbers in the above steps have been adjusted as development progresses. Please refer to the source code for accurate values.

 Mining Software

We expect there will be at least two mining methods at launch:

Use Mist client (official client implementation, based on Go language) for CPU mining.

GPU mining through a combination of the Ethereum daemon and sgminer. Data communication between the daemon and sgminer can be done through JSON-RPC
API completed. There are currently two APIs related to mining: eth_getWork and eth_submitWork.