ZK hardware acceleration aims to solve the problem of "inefficient ZK-SNARK proof generation".

Author: 雨中狂睡

Opportunities in blockchain often come from "solving market demands".

How to understand this? Let's take an example to illustrate.

During the previous cycle, the DeFi Summer on ETH was hot, but it also led to issues such as high ETH Gas fees and congestion. BSC emerged during that period, bringing many wealth opportunities. Apart from BSC, other well-known opportunities include Solana, Avalanche, and other Alt-Layer1 solutions.

Similarly, NFT Summer was the same. With the rise in ETH prices, the market urgently needed new wealth opportunities within the ecosystem to earn and accumulate ETH, or to leverage ETH, making NFTs the hot topic.

ZK hardware acceleration is the same. It aims to solve the problem of "inefficient ZK-SNARK proof generation".

At a previous conference, Vitalik mentioned that ZK-SNARK proof generation takes too long (about 20 minutes) and is inefficient. Ideally, real-time proof generation would be best. Therefore, to address this issue, Vitalik proposed three solutions: "parallelization and aggregation trees", "using SNARK algos and hash to improve efficiency", and "using ASIC for ZK hardware acceleration".

My understanding is that the first two solutions aim to improve ZKP efficiency from a technical upgrade perspective, which takes time for improvement (to be honest, I don't quite understand it, but can only grasp it superficially). Using "ASIC for ZK hardware acceleration" is like some people in the gaming world installing a spirit into their mouse, using external devices to solve the problem of low ZKP efficiency.

ZK hardware acceleration can be divided into GPU, FPGA, and ASIC. Why did Vitalik specifically mention ASIC? It's because from a flexibility perspective, GPU > FPGA > ASIC, and from a performance perspective, ASIC > FPGA > GPU.

In simple terms, GPU is general-purpose, ASIC is customized and specialized. General-purpose means everyone can use it, but the performance advantage is not as obvious, while ASIC chips can only adapt to a single solution. FPGA sits between ASIC and GPU, with a certain degree of flexibility, and its performance is stronger than GPU but weaker than ASIC.

Overall, ASIC can provide higher computing performance for ZKP because the current ZKP proof technology is relatively fixed. If ZKRollup seeks ZK hardware acceleration, GPU is a general solution, but in the future, customized high-performance ASIC chips are obviously the best solution.

Currently, the leading project for ZK hardware acceleration is the Cysic protocol, led by Polychain.

There is actually a lot of detailed information about Cysic. I am not from a technical background, so I can only share my understanding with everyone. Friends who are interested in the technology itself can visit the official website or media reports.

Cysic's hardware acceleration solution consists of two parts: ZKVM+ hardware design (GPU+ASIC). Currently, Cysic provides GPU acceleration solutions for ZKRollup (Cysic has reserved nearly 100,000 GPU devices, and its current advantage lies in GPU computing power, but its focus will shift to ASIC chips in the future).

ZKVM supports the virtual machine environment for ZK circuits, with the main advantage being the continuity and parallelism of ZKVM. In simple terms, when dealing with a large cake, high computing device, bandwidth, and memory requirements are needed. ZKVM allows us to cut the large cake into small pieces, making it more convenient to eat, with higher controllability and compatibility. Parallelism supports cutting and eating the cake at the same time, improving efficiency.

In terms of hardware design, Cysic packages the executor, which is used for computation, with a certain number of ZKVM chips and other necessary hardware into one chassis. This enhances the flexibility of the device in computation and its portability in the physical world.

In short, Cysic's design is based on considerations of cost-effectiveness and energy efficiency.

Finally, let me share my thoughts.

Why pay attention to the ZK hardware acceleration track? It's based on the optimistic view of the future development of ZKRollup. ZK hardware acceleration can help achieve real-time proof under external conditions, meeting the needs of ZKRollup. Therefore, when the ZKRollup narrative becomes popular, we can definitely look into ZK hardware-related projects, and there may be speculative opportunities at that time. Personally, I think it's not too late to speculate when the ZK narrative starts to heat up (which may be a bit difficult this year).

Reference:

"Why is Vitalik optimistic about ZK hardware acceleration?" https://www.odaily.news/post/5194489

https://twitter.com/BeWaterOfficial/status/1784083328737669165

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。