- Configure Benchmarks with the following:
- Instance Type (Available options: 8cpu-linux-x64, 16cpu-linux-x64, 32cpu-linux-x64)
- Programs (Comma-separated list of programs to benchmark)
- Benchmark Docker Image Tag
- Prover (sp1 or risc0)
- Use Cache (whether to use GitHub Actions Rust cache)
- Run Benchmark
- Start EC2 Instance (Handled by runs-on.com using compute-optimized instances)
- Setup Rust cache if enabled
- Pull Docker Image from registry
- Run Benchmark for each specified program
- Upload Results from Telemetry folder as GitHub artifact
Usage:
- The workflow can be triggered manually from the GitHub Actions tab
- Select the desired instance type from the dropdown (8, 16, or 32 CPU options)
- Provide the Docker image tag
- Enter the programs to benchmark as a comma-separated list
- Select the prover to use (sp1 or risc0)
- Choose whether to use Rust caching
- Results will be available as a downloadable artifact after the workflow completes
Available Programs:
- fibonacci
- rsa
- ecdsa
- json
- regex
- sha
- tendermint
- zkquiz
- bubble_sort
- iseven (sp1 only)