TLDR

  • Buterin states that the state tree and VM account for more than 80% of proving expenses

  • EIP 7864 proposes swapping the hexary tree for a binary structure

  • Reduced Merkle proof length could lower bandwidth consumption and gas usage

  • RISC V VM suggested as a long-term substitute for the EVM

The upcoming phase of Ethereum scaling might prioritize its foundation over Layer 2 solutions. To tackle proving efficiency, Vitalik Buterin has suggested significant changes to the execution layer. He pinpointed the state tree and virtual machine as the main bottlenecks.

In recent updates, Vitalik Buterin noted that these elements are responsible for more than 80 percent of proving expenses. He contended that structural reform is necessary for efficient zero-knowledge proving. His strategy encompasses redesigning the state tree and implementing long-term VM modifications.

Vitalik Buterin’s Binary Tree Proposal via EIP 7864

EIP 7864 is central to the state overhaul. This proposal intends to substitute Ethereum’s existing hexary Merkle Patricia Tree with a binary tree. The objective is to minimize proof size and enhance verification efficiency.

Binary trees are expected to generate Merkle branches that are roughly four times shorter. This decrease alleviates bandwidth requirements for light clients and privacy tools. It could also cut costs for systems like Helios and private information retrieval.

Buterin explained that the new design would organize storage slots into pages. Each page would accommodate 64 to 256 slots, equating to approximately 2 to 8 kilobytes. This arrangement enables more efficient loading and updating of related storage.

He observed that numerous decentralized applications access storage slots that are adjacent. With the new model, these patterns might save more than 10,000 gas per transaction. Additionally, the redesign lessens the variance in access depth between small and large contracts.

The proposal involves swapping keccak for more efficient hash functions. Mentioned alternatives include blake3 or Poseidon variants. These modifications seek to boost proving speed and render the state tree more prover-friendly.

Vectorized Math Precompile for Short-Term Benefits

Prior to major VM modifications, a short-term upgrade is proposed. He recommended incorporating a vectorized math precompile into the Ethereum Virtual Machine, describing it as “the GPU for the EVM.”

This precompile would facilitate 32-bit or 64-bit operations on number lists simultaneously. It has the potential to speed up cryptographic tasks like hashes and STARK validation. Buterin indicated this could enhance performance by a factor of 8 to 64.

He also connected this upgrade to future quantum-resistant signatures and fully homomorphic encryption. The proposal aims to diminish dependence on complex workarounds and to refine raw execution within the existing framework.

Long-Term Vision to Transition Away from the EVM

Buterin also detailed a long-term strategy to replace the Virtual Machine. He proposed shifting to a RISC V-based architecture, arguing that many provers already function within RISC V environments.

He stated that Ethereum has increased in complexity because of special-case precompiles. He noted that developers frequently bypass direct use of the EVM. In Buterin’s view, a simpler, more efficient VM would align better with Ethereum’s general-purpose design.

The roadmap suggests a phased deployment approach. RISC V could initially drive precompiles. Subsequently, users might deploy contracts using the new VM. Ultimately, the EVM could function as a compatibility layer constructed within the new architecture.

Buterin remarked that Ethereum would operate sufficiently with incremental upgrades. However, he firmly believes that the necessity for deeper changes will become apparent as state reforms mature. His proposal redirects focus from rollup capacity to Ethereum’s core execution design.