Armoni FlowContact2,151 bytes of EIP-170 headroom
Web3 Lab / Smart Solve
Numerical computation, measured for the EVM.
An experimental Solidity framework for evaluating the precision, gas cost, and feasibility of numerical algorithms on Ethereum-style execution environments.
0x01
Solidity
EVM benchmarks
Framework
What Smart Solve is
Smart Solve is a numerical-computation and benchmarking framework written for Solidity. It is a Diamond-based production-path demonstrator built to study the practical constraints behind on-chain numerical work: gas, precision, runtime bytecode, deployment, and execution behavior.
Scope
What it can do
The current benchmark suite covers a range of small, bounded numerical workloads.
01
Root finding
02
Numerical integration
03
Differentiation
04
ODE solving
05
Polynomial evaluation
06
Matrix operations
07
Linear solvers
08
Optimization / steepest descent
Measured results
Benchmark highlights
Selected Diamond-routed measurements from the project. Figures below are intentionally limited to recorded results.
| Benchmark | Result | Configuration |
|---|---|---|
| Full SmartSolve Diamond deployment / install | ≈ 25,704,269 gas | Complete deployment path |
| Bisection root finding | 1,125,674 gas | 42 iterations |
| Simpson 1/3 integration | 80,327 gas | n = 2 |
| Centered-difference differentiation | 62,393 gas | Diamond-routed |
| Euler ODE solve | 53,837 gas | 3 steps |
| Horner polynomial evaluation | 32,319 gas | Diamond-routed |
| 2×2 matrix multiplication | 68,996 gas | Diamond-routed |
| 2×2 Gaussian elimination | 105,374 gas | Diamond-routed |
| Steepest descent optimization | 210,385 gas | Diamond-routed |
down from 154,956 gas on the Diamond route
less precise in the 1/3 rounding case
Method
Two ways to read a benchmark
The framework separates algorithm behavior from the cost of a production-style proxy and facet call path.
| Execution path | What it measures | Included overhead |
|---|---|---|
| Harness / direct | The algorithm itself in isolation. | No Diamond selector lookup or delegatecall. |
| Facet-direct | A facet’s implementation path. | Facet execution without the Diamond fallback route. |
| Diamond-routed | Production proxy / facet behavior. | Fallback selector lookup and delegatecall overhead. |
What this is / what this is not
What this is
- Experimental Web3 numerical computation framework
- Gas and precision benchmarking framework
- Solidity / EVM feasibility study platform
- Engineering proof-of-work for advanced smart contract systems
What this is not
- An audited production DeFi math library
- A universal replacement for off-chain computation
- A claim that all numerical algorithms are economical on mainnet
- A guarantee that every benchmark measures full binary128 empirical accuracy
Applications
Where it is useful
Feasibility decisions
Web3 teams evaluating whether a numerical method can responsibly run on-chain.
Bounded DeFi prototypes
Small, constrained numerical workloads for oracle, risk-model, or DeFi experiments.
Optimization studies
Gas and bytecode work for Solidity contracts with math-heavy execution paths.
Research and due diligence
Academic EVM research and a concrete technical example for smart-contract review.
Constraints
Gas-cost feasibility has limits
- Some workloads are only practical at small, bounded sizes.
- Large matrix and iterative workloads can approach or exceed the 30M gas limit.
- Heavy scalability tests are opt-in rather than assumed safe defaults.
- Callback-driven algorithms include callback staticcall cost in their gas measurement.
- Results distinguish harness/direct, facet-direct, and Diamond-routed execution.
Evidence
Engineering quality and reproducibility
- Structured benchmark JSON exports
- Reproducibility metadata
- Bytecode budget guardrails
- Deployment artifacts and post-deploy verification
- Failure and out-of-gas classification
- Precision-aware tests where available
Smart Solve
Explore the implementation, benchmark artifacts, and experiment design in the public repository.
View on GitHub