Is DeepBrain Chain, the high-performance computing track, just needed in the Web3.0 era?
In recent years, with the increasing maturity of Internet infrastructure, fields such as big data analysis, Internet of Everything, and artificial intelligence have shown a geometric growth trend and developed rapidly. According to statistics, the scale of China's artificial intelligence market has exceeded 400 billion yuan in 2020, with an average compound growth rate of 40%.
The development speed of foreign institutions cannot be underestimated. The top AI organization Tractica predicts that the hardware and software foundation driven by artificial intelligence The total global market for facilities is expected to grow to USD 115.4 billion by 2025. Among them, AI cloud computing infrastructure will account for 70%.
Yes, what supports all of this is massive computing power. The traditional ultra-large data center can no longer meet the market demand, and it is necessary to schedule and manage different levels of data centers (cloud computing nodes, metropolitan nodes, edge nodes). The market needs a low-cost, high-performance unified decentralized computing network solution .
Decentralized high-performance computing network DBC
Traditional cloud computing consists of three parts: computing, storage, and CDN (content delivery network). The leader in the storage track is Filecoin, and the veteran Arweave and Storj are also doing well. Looking at the CDN field again, Theta is thriving. In terms of computing, DFINITY, which has just launched its mainnet recently, has occupied the territory of CPU computing, and there are many representative projects in the field of GPU computing, and DeepBrain Chain is one of them.
Compared with DFINITY’s CPU computing, GPU is more focused on high-performance computing, especially for the field of artificial intelligence. The deep learning algorithm behind it fits well with GPU computing. Whether it is AlphaGo, which plays games with top human chess masters, or the intelligent robot "Xiaodu" who accompanies us every day, it seems that a scene full of artificial intelligence and everything can be interconnected is close at hand. In addition, many scenarios such as autonomous driving, biopharmaceuticals, artificial intelligence, cloud games, and blockchain require the support of high-performance computing power. Last year, Nvidia's market value surpassed Intel's, which also marked the mainstream trend of traditional cloud computing shifting from CPU to GPU. Especially for the visual rendering track, the current global annual market is more than 200,000 GPUs.
DeepBrain Chain technology has realized a global decentralized GPU high-performance computing power scheduling network based on blockchain technology . DBChain is an open source GPU computing power pool and GPU cloud platform. Idle GPU computing power resources around the world can freely join the DeepBrain Chain computing power network. Through network coordination and deployment, the computing power in the domain will be utilized most efficiently and form a A complete ecological closed loop. The core of DeepBrain Chain is to solve the problem of expensive artificial intelligence computing power, and realize the unlimited expansion of massive computing power, so as to promote the popularization and democratization of artificial intelligence computing power, and accelerate the advancement of human society towards intelligence.
There are currently more than 6,000 GPUs in the DBC network, and with the start of computing power ecological construction, nodes in the network The number will increase. There are 400 million DBC computing power ecological construction rewards every year to motivate high-performance computing power providers and attract more GPU resources into the ecology.
When ETH2.0 goes online in the future, the network consensus protocol will change from PoW to PoS mechanism. There is a high probability that the graphics card in the computer will become an idle resource. Among them, high-end graphics cards may also participate in the ecological construction of computing power in DeepBrain Chain.
From NEO to Substrate application chain
Previously, DeepBrain Network was established in the NEO ecosystem and implemented with smart contracts. However, the limitations of smart contracts make it difficult to implement many complex functions. Now, DeepBrain has transferred to the Polkadot ecology.
From a security point of view, the Polkadot network implements a shared security model, and the Substrate chain wins the parachain slot Next, you can join the Polkadot system to share the security of the relay chain.
Substrate application chain, as the underlying development framework of DeepBrain Chain, can realize the complete customization of DeepBrain The business logic of the network is implemented in depth. At the same time, the Substrate application chain has endowed DeepBrain Network with the ability to interoperate with other application chains through IBC and XCMP. At the same time, the Polkadot ecological transfer bridge facility has also been prepared, which enables DeepBrain Chain to interact with other existing public chains . In addition, Polkadot's governance mechanism is relatively complete, and its ecology is huge, and it has integration points with DeepBrain's business. The combination of protocols will produce unpredictable synergies.
The stability of the bottom layer brings about the development of ecology, in blockchain, artificial intelligence, cloud games, visual rendering , biopharmaceuticals, semiconductor simulation and other scenarios have been widely used, providing cost-effective GPU computing power for many enterprises. A total of nearly 50 manufacturers (developers) around the world have deployed high-performance GPU cloud platforms based on the DeepBrain Chain network, serving Hundreds of companies and tens of thousands of AI developers.
Governance of decentralized computing power network
DeepBrain Chain will move toward decentralized governance after the mainnet is launched. The opinions of members of the community will play a more important role in the DeepBrain Protocol. In the future, whether it is technology development, marketing promotion, development strategies, etc. All will be governed through the DeepBrain Chain decentralized governance platform. Anyone who holds DBC tokens has the right to make a proposal. The proposal can be anything. Whether the proposal is implemented will be automatically completed through community voting.
For example, DBC holder A is a technology developer and thinks that function X is very important to DeepBrain Chain, and he Have the development ability to complete the development of this function. At this time, technical staff will initiate a proposal, which will describe in detail the importance of functional requirements to DeepBrain Chain, and then initiate a vote in the community to obtain development funds. In the end, all community members holding DBC tokens can participate in voting and decide whether to appropriate this proposal.
As another example, any mechanism improvement or major product function deployment related to DBC network development will also be decided by community vote . All modifications to the DBC protocol will be done in a decentralized manner.
The ultimate goal of the DeepBrain Chain Foundation is to strive to complete the complete removal of DBC governance after the first year of the mainnet launch centralized. In the future, participating in the governance on the DBC chain will be a very good way for individuals to start a business. The DBC community has created a fair and free platform for all entrepreneurs in the Web3.0 era.
DBC Token Economic Model
The value of DBC tokens is related to the number of leased GPUs in the DBC network. The more leased GPUs, the greater the value of DBC.
The destruction ratio of each GPU server rental transaction amount of the user is 100%. For high-performance computing power providers, each graphics card needs to pledge a certain amount of DBC and lock it. Take the upcoming Computing Galaxy Competition as an example. In the Galaxy Competition, each graphics card is pledged with 100,000 DBC, which needs to be pledged for at least 365 days; when the machine has been online for more than 365 days and no users have used it for more than 10 days, you can withdraw from the pledge to stop computing power For ecological construction, the pledged DBC will be released within 3 days, and the previously obtained rewards will continue to be released linearly.
When the number of GPUs pledged by graphics cards on the entire network exceeds 10,000 cards, the number of DBC pledges for new graphics cards will be automatically adjusted, according to the flow DBC quantity and DBC price adjustments. Starting from 10,001 cards, the formula for the total number of pledges per card is: 100,000 × 10,000/n; if calculated according to the above formula, the pledge of a single card exceeds 50,000 yuan, then the pledge will be calculated at a maximum of 50,000 yuan.
If there are 5000 cards in the whole network, 500 million DBC will be locked, and 500 DBC will be locked every month The rent of RMB 10,000 is used to destroy DBC. According to the DBC price of 0.1 yuan, 50 million DBC will be destroyed every month, and more than 600 million DBC will be destroyed every year; in the process of actually completing the pledge and destruction, the DBC price is Floating, the higher the DBC price, the less and less the number of DBC destroyed by a single card in a single month.
This token economic model not only ensures the deflation trend of the total amount of DBC as a value support, but also ensures the Stability to ensure the sustainable development of the DBC network.
In order to give back to the early GPU computing power contributors of DeepBrain Chain, DeepBrain Chain is expected to be held on June 2021, Beijing time. After the 23rd, the galaxy computing power competition will start. There are no restrictions on the competition, as long as the graphics cards meet the requirements, any provider of high-performance computing power can participate. After 10:00 on June 23, at any time, as long as the main network participates in the pledge of the number of computing power ecological construction machine graphics cards greater than or equal to 5,000, the competition will start immediately. This Galaxy Hash Power Contest will last for 1 month. For the top 5 computing power pools, the DBC team will take out an additional 30 million DBCs as rewards, which will be distributed according to the percentage of computing power.
There are also requirements for participating GPUs, the following are the minimum requirements for participating GPU servers:
1) GPU: 2 cards or 4 cards; NVIDIA Nvida series graphics card, the minimum requirement is 8G video memory; considering the rental rate, the recommended graphics card type is 2080ti, 3060ti, 3070, 3070ti, 3080, 3080ti, 3090, etc.;
2) Memory: 448 G (4 cards) or 224 G (2 cards), (calculation formula: number of GPUs*112);
3) CPU: 2 cards - 1 CPU, 4 cards - 2 CPUs, Intel, V3 or above, main frequency 2.5Ghz, turbo frequency 3.0Ghz and above, single CPU 10 cores and above;
4) Hard disk: system disk 512g (SSD), data disk 2T (mechanical hard disk);
5) network Bandwidth: 2 cards-10M/set or more; 4 cards-20M/set or more;
6) Independent fixed IP;
7) The machines are placed above the T3 level of the standard data center to maintain the stability of the network and electricity.
Only GPU servers that meet the minimum requirements are eligible for rewards. A single GPU computing power point has two impact indicators: the number of GPU CUDA Cores and the size of video memory. The number of GPU CUDA Cores affects 70%, and the impact of video memory accounts for 30%.
It should be noted that the number of DBC online rewards is based on the share of the computing power provider in the total DBC network computing power 4 elements that affect DBC online rewards: GPU computing power value, GPU online number, GPU online number under a single DBC wallet, GPU rental status;
a. GPU computing power calculation formula:
Computing power=13+0.7×√(cuda number)+0.3 ×Video memory/24×√(number of cuda)
In the above formula, √ means the square root, and the reference value of some graphics cards is as follows:
b . The online number of GPUs under a single DBC wallet, for each additional GPU, the computing power of all machines will increase by 0.01%, until the increase is 10%;
c. Online rental status: when the GPU is rented, the computing power value increases by 30%;
Just like a hackathon in the DeFi ecosystem, Galaxy Competition not only allows players to understand DeepBrain Chain, but also selects loyal users for the project.
As a project that crosses bull and bear on the decentralized computing track, DeepBrain Chain is realizing large-scale implementation. With the launch of the mainnet, let us wait and see how DeepBrain Chain will perform in the Web3.0 world.
Welcome to join the official BlockBeats community:
Telegram Subscription Group: https://t.me/theblockbeats
Telegram Discussion Group: https://t.me/BlockBeats_App
Official Twitter Account: https://twitter.com/BlockBeatsAsia
