Choose the job before the number.
The IR balanced build cut setup gas by 8.60% for the counter, 5.43% for escrow and 12.68% for the fee-hook fixture, including its factory. It increased setup gas for player stats by 4.81%.
BUILD COST LAB / UNCHANGED APPLICATION CODE
Same code.
Smarter build.
A smaller contract can be cheaper to set up. A larger one can be cheaper to use. Compare four ways to build the same app before choosing which trade-off fits.
Gas measures the work a blockchain transaction uses. Less gas can mean a lower fee at the same gas price.
4 APPS · 4 BUILD SETTINGS48 FRESH RUNS · 744 CHECKSNO SOURCE CODE CHANGES
CHOOSE AN APP. FOLLOW THE COST.
Player stats
The estimate uses measured setup and transaction gas. It is not a prediction of future fees.
LOWEST COST IN THIS CALCULATION
Comparing builds…
—
Choose a measured action.
Every build uses the same application source. The comparison checks returned values, permissions, rejected calls, events and tested state changes.
Loading measurements…
| BUILD | SETUP GAS | GAS / ACTION | SETUP + ACTIONS | CONTRACT SIZE | CHECKS / RUN |
|---|
Studio currently builds with Standard settings. Open the downloaded test project to run this comparison’s full test plan with that build.
THE TRADE-OFFS MATTER
More optimization.
Not always less cost.
Frequent use has to earn its keep.
The IR frequent-use build made player-stat setup 46.20% more expensive and updates 3.42% cheaper. A lighter build can be the better choice when an app will make only a few updates.
These results compare compiler settings on four fixed examples. The earlier 75–83% storage savings came from changing the data layout. They are a separate experiment, not a compiler-wide promise.
DETAILS FOR BUILDERS
Every setting.
Every result.
What are the four builds?
All use Solidity 0.8.30, the Paris EVM target, optimizer enabled and identical application source. Standard uses the legacy code generator with optimizer-runs 200. Smaller build uses runs 1. The two IR builds use the Yul intermediate pipeline with runs 200 or 10,000. “Runs” is a compiler cost preference; it is not the number of optimization passes or transactions executed.
How were costs and behavior checked?
Each of 16 compiled artifacts ran on 3 fresh, isolated Anvil 1.8.6 Cancun chains. All 48 runs passed their scenarios, totaling 744 checks. Gas comes from transaction receipts, including base and input-data costs. Deployed code matched its compiler artifact outside declared immutable references. ABI, storage layout, tested outputs, reverts, event data and escrow value movements were compared. This is functional evidence, not an independent security audit or proof for every input.
The fee rule runs at a valid CREATE2 hook address and tests the beforeSwap callback. It does not run a complete PoolManager swap or liquidity lifecycle. Setup includes its factory and valid hook deployment, excluding the deliberately rejected constructor test.
Can I switch every contract to IR?
IR has documented edge-case semantic differences. Test the actual contract and required EVM target before changing a build. No new pipeline has been selected automatically for your Studio projects.
Where can I inspect or reproduce this?
The download contains all original build and execution packets, sources, test plans, compiler settings, hash manifests, a summary analyzer and the bounded native-run script. Running the native experiment needs the separately pinned Linux compiler/runtime already used by Studio; it installs nothing.
Current application source SHA-256: