- Dividing Tasks into 'Glue' and 'Coprocessor'
- Practical Examples in Ethereum and AI
- Future of Cryptography and New Approaches
Ethereum creator Vitalik Buterin has shared a new concept for organizing computations by splitting them into two parts: 'glue' and 'coprocessor'. This article provides an explanation and examples of this model.
Dividing Tasks into 'Glue' and 'Coprocessor'
The concept proposed by Vitalik Buterin suggests dividing computational tasks into two types. The 'glue' handles general, less intensive tasks, while the 'coprocessor' takes care of heavy, structured computations. Vitalik explains that most computations in systems like the Ethereum Virtual Machine (EVM) are already divided this way. Some parts of the process require high efficiency, while others are more flexible but less efficient.
Practical Examples in Ethereum and AI
A recent Ethereum transaction illustrates how gas is allocated across different tasks. In a transaction where Vitalik updated his blog's IPFS hash on the Ethereum Name Service (ENS), the total gas consumption was 46,924. Key resource-intensive operations, including storage reads and writes, cryptography, and logging, took about 85% of the total gas. In artificial intelligence (AI), a similar pattern can be seen, where most work is done through vectorized operations, while high-level logic remains in Python.
Future of Cryptography and New Approaches
Vitalik believes this approach is becoming more common in modern cryptography, such as in SNARK proofs and multi-party computation (MPC). Developers create specialized modules for the most costly operations and combine them with general proving systems, achieving both efficiency and flexibility. This model inspires the creation of chips optimized for security and open-source design, using proprietary ASIC modules for more intensive computations.
Vitalik Buterin emphasizes the importance of balancing general and specialized tasks in computations. The 'glue' and 'coprocessor' approach may be key to building more efficient and secure systems, both in blockchain and other fields.
Comments