Why can’t more than 95% of mining machine manufacturers mine Filecoin?

In the previous sharing, I mentioned that from the time when the Filecoin testnet was launched and the official hashrate rankings were released to the time when the hashrate rankings were later removed, during this period, according to our observation, only less than 10 companies’ nodes were included in the hashrate rankings. Although this hashrate ranking can only be used as a reference and not as a final criterion, we believe that if a company cannot even make it into the hashrate rankings, it may be difficult to mine after the mainnet is launched.

Does that mean that only the companies on the list can mine? Not necessarily, but in any case, there won’t be too many companies that can mine. And how many nodes are participating in the test network? There are at least more than 200. There are even more companies on the market that produce mining machines under the banner of "Filecoin mining".

There are so many mining companies, but only a few mining machines can mine. According to the current proportion, less than 5% of the companies can mine. In other words, it is possible that 95% or even a higher proportion of mining machine companies produce mining machines that cannot mine, which will undoubtedly reshuffle the industry.

We have predicted this situation before, and believe that:

First, in our opinion, Filecoin mining will eventually be monopolized by a few large mining sites like Bitcoin mining, and the vast majority of mining companies will be ruthlessly eliminated.

Secondly, with so many companies on the market collecting money from customers but unable to mine, they may have to privately seek technical cooperation with companies that can mine to help themselves resolve this embarrassing situation.

Why is this so? Filecoin mining is not something that can be mined by just any computer with a hard drive connected to the Internet. It involves a very high level of technology.

Before the testnet was launched, an official statement revealed the core and key technology here. This statement was the "GPU required for mining" statement that caused a stir in the community. While introducing the GPU, this statement also gave a detailed mining process, especially the calculation process of the proof of replication and the proof of time and space.

Here we analyze the calculation process of these two proofs.

First, let's look at Proof of Replication. For Proof of Replication (PoRep), the system's calculation is divided into four steps:

The first step is to set the parameters. This step does not consume computing power and does not use the GPU.

The second step is to copy the data to form a specific code. This step is time-consuming and consumes CPU resources, but the GPU cannot be used, so the GPU will not be used.

The third step is to generate a "proof of replication", which requires the calculation of "zero-knowledge proof". This process can be accelerated by using GPU.

The fourth step is to self-verify and generate the proof, which does not consume computing power and does not require the use of GPU.

Regarding the calculation of "Proof of Spacetime", the Filecoin project has recently added the "election post" indicator. This indicator requires miners to quickly calculate the "Proof of Spacetime (PoSt)" in less than one block cycle after mining a block. At this time, GPUs are more important.

So in summary, in the whole process, whether or not GPU is used, the whole calculation process is quite complicated, which is very different from our traditional understanding that the decisive factor of hard disk mining lies in the read and write speed of the disk. In Filecoin mining, the calculation process of these two proofs is the key.

So how do the two popular frameworks of cloud computing and stand-alone mining machines handle this calculation?

As for cloud computing power, since it uses a cluster mode, there is no single mining machine in the system but only functional modules. Therefore, the whole system is like building blocks, with different building blocks responsible for different functions. If a certain function is weak, just add a module. The overall architecture is relatively flexible, with strong scalability and high risk resistance.

For a single mining machine, it is actually a "cluster operation" to some extent. However, the cluster operation here adopts the method of connecting mining machines in series, connecting several mining machines in series to form a large node to participate in mining.

No matter which of these two architectures is adopted, it requires very high technical content. For cloud computing companies, the core technology lies in the design of the overall framework and the overall system; for single mining machine companies, the core technology lies in the series connection of mining machines.

Therefore, the vast majority of companies on the market that want to make a fortune by taking advantage of the Filecoin hype do not have the technical capabilities at all, so it is impossible for them to mine.