Solidity Gas Optimisation: Storage, Packing, Assembly

Frequently Asked Questions

01.What is Solidity storage slot packing and why does it reduce gas costs?

Solidity storage slot packing is the compiler behaviour where multiple variables smaller than 32 bytes are placed into a single 32-byte storage slot. Each SSTORE or SLOAD operation costs gas per slot, not per variable, so fitting two uint128 values into one slot halves the I/O cost compared to two separate uint256 slots. Engineers control packing by ordering struct fields from smallest to largest type, or grouping sub-256-bit types consecutively. Correcting a four-field struct from naive to packed order can reduce cold SSTORE cost from 80,000 gas to 40,000 gas, with no assembly required and no change to the external ABI.

02.How does EIP-2929 affect SLOAD and SSTORE gas costs in production contracts?

EIP-2929, activated at the Berlin hard fork in April 2021, introduced the concept of cold and warm storage access. A cold SLOAD now costs 2,100 gas and a cold SSTORE costs 20,000 gas for a zero-to-nonzero write. Once a slot is accessed within the same transaction it becomes warm, reducing subsequent SLOAD to 100 gas. Engineers can pre-warm slots using EIP-2930 access lists included in the transaction envelope, converting the first access from cold to warm pricing at a cost of 1,900 gas per storage key in the list. Break-even point is one cold SLOAD avoided per key.

03.When should a Solidity engineer use inline Yul assembly instead of high-level Solidity for gas savings?

Inline Yul assembly is justified when high-level Solidity generates demonstrably suboptimal opcodes on hot-path functions: tight loops over packed storage, bulk bitwise operations on packed booleans, or custom memory layouts for calldata decoding. The cost is auditor complexity and loss of static-analysis coverage. Ancilar applies Yul only after profiling shows measurable gas reduction on a representative transaction set, and only in functions that carry explicit NatSpec safety annotations. Any Yul block must be covered by invariant fuzz tests and reviewed as a separate, documented audit scope before deployment.