gas53

Optimizing Gas Consumption in Ethereum Smart Contracts

The study introduces four gas-waste detectors integrated into the Slither static-analysis framework and applies them to 397 Solidity smart contracts from 53 Ethereum dApps, identifying 250 gas-waste instances.

Dataset

Download verified-gas-opt-ppr.zip

• 53 dApps deployed on Ethereum Mainnet
• 397 Solidity smart contracts
• Each dApp is organized by its Ethereum contract address
• Each folder contains Solidity source files and corresponding analysis reports

Gas-Waste Detectors

The dataset is annotated using four detectors introduced in the paper:

• cache-address-this — repeated use of address(this)
• div-by-uint — division by unsigned integers
• smaller-string — constant strings of 32 characters or fewer
• unnec-casting-same-type — unnecessary type casting to the same type

Key Findings

• All 53 dApps contained at least one of the analyzed gas-waste patterns.
• More than 50% of detected gas-waste instances involved division by unsigned integers.
• More than 90% of the dApps exhibited division-by-unsigned-integer gas inefficiencies.
• Approximately half of the dApps contained short constant strings that could potentially be optimized.
• Unnecessary same-type casting was the least common detected gas-waste pattern.

Publication

Optimizing Gas Consumption in Ethereum Smart Contracts: A Static Analysis Approach

Abdur Rehman Raza, Muhammad Hanzala, Khawir Mahmood, Shahzaib Tahir, Saifullah Khalid, and Asif Masood

2025 International Conference on Communication Technologies (ComTech), pp. 1–6, IEEE, 2025.

View the published paper

Citation

If you use the Gas53 dataset or its analysis results in academic work, please cite:

@inproceedings{raza2025optimizing,
  title     = {Optimizing Gas Consumption in Ethereum Smart Contracts: A Static Analysis Approach},
  author    = {Raza, Abdur Rehman and Hanzala, Muhammad and Mahmood, Khawir and Tahir, Shahzaib and Khalid, Saifullah and Masood, Asif},
  booktitle = {2025 International Conference on Communication Technologies (ComTech)},
  pages     = {1--6},
  year      = {2025},
  publisher = {IEEE},
  doi       = {10.1109/ComTech65062.2025.11034442}
}

This site is open source. Improve this page.