Parallel Execution

Traditional EVM chains process transactions one at a time in strict sequential order. Parallel execution identifies transactions that do not access overlapping state (e.g., two unrelated token transfers) and processes them simultaneously across multiple CPU cores.

The key challenge is detecting conflicts. Optimistic parallelism, used by Monad and Sei V2, speculatively executes transactions in parallel and re-executes any that conflict. Block-STM, pioneered by Aptos (and adapted from research at Meta's Diem project), uses a multi-version data structure to detect conflicts efficiently. Solana uses a declared access list approach where transactions specify which accounts they touch, enabling the scheduler to parallelize safely.

Monad, one of the most anticipated parallel EVM projects, claims 10,000 TPS in internal benchmarks with full EVM bytecode compatibility. According to their technical documentation, Monad combines optimistic parallel execution with a custom asynchronous I/O layer for state access. Sei V2 launched its parallel EVM upgrade in mid-2024, achieving roughly 5x throughput improvement over sequential execution according to Sei Foundation benchmarks.

For investors, parallel execution represents the next scalability leap for EVM-compatible chains. It preserves developer tooling and smart contract compatibility while dramatically increasing performance — a compelling value proposition versus building entirely new execution environments.