The AMA (Ask Me Anything) of Qitmeer Mining was hold on 3rd August online introduced by the mining team leader Mr. Eric who has 16 years software development and 4 years blockchain development experience.
Q1: Why does Qitmeer choose POW instead of POS or DPOS?
Eric: There are various consensus like POS, DPOS, PBFT etc., however we believe that POW is the most decentralized consensus that provides high security.
There are three main categories of consensus in blockchain. The first one an open network where participants can join and leave the network freely without cost which is represented by POW that is the most flexible and decentralized choice. The second one is authorizing the right to validate blocks to the communities, such as holding the stake or tokens represented by POS. The third one is an enclosing way that the person who has right to validate blocks should holds huge amount of stake or stakeholders who will reach a consensus represented by DPOS, PBFT or BFT. That is to say that if there’s a condition to join the network such as holding stake, we can define as POS, furthermore that if only a certain group of people can join, we can define as DPOS.
Thus, we can see that consensus becomes less decentralized from pow to pos, then to DPOS. Whereby we choose POW as it shares the same value proposition of Qitmeer.
Q2: Can you introduce some mainstream ASIC-resistant algorithms?
Eric: There are four mainly types of ASIC-resistant algorithms.
The first type is Memory Intensive Algorithm. This type of algorithm requires large memory and memory bandwidth which meet the characters of GPU, like EquiHash (Zcash) and Ethash (Ethereum). Thus, it is GPU friendly at the beginning.
The second type is RandomHash Algorithm which takes the advantages of the randomness of algorithm, like the X16R of Ravecoin which is improved to X16S. The execution sequence of several algorithm in POW will be decided by the several latter bits of last block randomly which makes its resistance to ASIC which has to input a fix algorithm to the chip.
The third type is mixed algorithm which is not completely ASIC-resistant represented by Grin. There two algorithms in Grin, one is Cuckaroo which is ASIC-resistant where the other is Cuckatoo which is ASIC-friendly. A algorithm mechanism written in the code will adjust the speed of producing new blocks of the two algorithms according to the situation of network.
The last one is forcible hard fork, like Monero. Their team modified its algorithm by a hard fork to improve ASIC-resistant as some manufactures input their previous algorithm to mining machines.
There are some pros and cons of ASIC. ASIC machine is more secured, stable and effective than CPU and GUP, however, it has higher cost and longer development period where CPU and GUP are more flexible with lower cost and threshold.
However, the security of CUP mining will be lower than ASIC when the computing power is not large enough at the beginning stage. ASIC mining machine is specially designed for the public chain as the partner and stakeholder, whereby everyone will cooperate to maintain the security for the same objectives of a sustainable development. As CPU and GPU are common hardware devices which are easily to be collected to form a big computing power which may be enough to attack the public chain technically, but it may only happen when there’s huge benefits. Hence, most of the project will prepare some computing power at the beginning stage if they accept GPU or CPU mining.
Q3: Why do Qitmeer choose Cuckoo Cycle as our ASIC-resistant algorithm?
Eric: This decision is made after a research as the design philosophy of Cuckoo cycle is close to our project which can be concluded to 2 mainly targets as follows.
Fair mining. We hope that the effectiveness of mining shouldn’t have too much difference between mobile devices and household appliances like cellphone and TV box.
Energy and environment friendly. This algorithm is designed to make use of memory space instead of large computing which can save more energy.
Qitmmer proposes to become an energy and environment friendly public chain with the achievement of fairness, and Cuckoo cycle is based on graphs where DAG (Directed Acyclic Graph) which is one of characters of our project. Therefore, we choose Cuckoo cycle as our ASIC-resistant algorithm.
Q4: Which mining device suits Qitmeer network?
Eric: Qitmeer is GPU friendly. A GUP with medium configuration can mine in Qitmeer.
According to current test, GUP with more than 5 GB memory is qualified. There are two types of GPU, NVIDIA and AMD. NVIDIA is more effective than AMD whish was tested under the calculations of 24 edge bits.
Due to the calculation of Cuckoo cycle is based on graphs, GPS (graph per second) is the unit of mining efficiency. The result we have tested that the GPS of NVIDIA 1060, 1070 and 1070pi are 4.4, 4.9 and 5.3 separately, while the GPS of AMD 580 and 590 are 2.7 and 3. From the mining efficiency perspective, AMD is better than NVIDIA.
Miners should consider the actual hardware input cost and power consumption comprehensively as electricity is a long-term investment.
Q5: What’s the status and plan of the development of mining?
Eric: Currently, we have deployed Cuckoo cycle POW to Qitmeer and the mining procedures based on GPU has been open source since 31st July, an internal mining test is going on.
Two targets are planning to meet in the future, one is to keep improving the execution efficiency of GPU mining; the other is to consider and conduct research on mixed mining model. We might introduce another one or several ASIC-friendly algorithms together with Cuckoo cycle POW. By adjusting the speed of producing new blocks of various algorithm to control of the algorithm of code level, we hope to have more miners to join the network. We would like to have a desired balance of the outcome of mining between GPU and ASIC.
Q6: Please explain the philosophy of fair mechanism and the execution of the set plan.
Eric: The fair mining is mainly reflected in two aspects.
Firstly, fair mining devices. The purchase of mining devices shall be relative fair and between hardware manufacturers, big institutions, big mining pool and individuals and a relative low threshold which will enable more people to be involved easily.
Secondly, fair mining output. It is mainly reflected in the fair input and output whether it is GPU or ASIC mining machine, the hardware cost and electricity cost you invest should have a positive correlation with your mining income.
Q7: What will be the output of mining in the testnet? Can Pmeer be cashed out? What’s the relationship between Pmeer, Meer and HLC?
Eric: Pmeer is the output of testnet mining. The relationship between Pmeer, Meer and HLC is related to economic model that all parties’ benefits will get a balance which will be disclosed later in September.
Q8: Should the ASIC mining machine be customized if Qitmeer has a mixed mining algorithm?
Eric: Yes. Because of the uniqueness of our block structure and algorithm, the available ASIC mining machine in the market can’t mine Meer without change the algorithm in the hardware. Although there are some similarities between Qitmeer and Grin, still some details and block structure are different, also there are difference with bitcoin and other public chains.
Q9: Why Qitmeer doesn’t choose CPU mining?
Eric: Unlike ASIC, CPU is a kind of multi-purpose computer hardware equipment. It will be more complicated if project designs a few CPU friendly but ASIC resistant.
No public chain so far has a long-time-proved application developed using CPU friendly algorithm. Of course, there are some project accepts it like RandomX of Monero, which was described in their whitepaper that “Practically everyone in the world now has a smartphone in their pocket with a CPU and memory that’s capable of mining RandomX”.
But think about this. If a CPU friendly mining algorithm like RandomX is implemented, laptops, mobile phones, Ipads and so on can all mine, will it be controlled by hackers and lead to a large number of botnet mining? These problems are worth studying, and we will continue to study on RandomX.