-
Notifications
You must be signed in to change notification settings - Fork 39
Open
Labels
P0-🔥ASAPASAP
Description
Product Perspective
EVM developers expect fast test execution. The standard Hardhat local testing experience completes
in ~1 second. While Hedera's consensus model inherently adds latency, Solo must minimize this gap
to provide a developer experience that doesn't frustrate users migrating from Ethereum tooling.
Local Node's turbo mode provides accelerated test execution. Solo must offer equivalent
capabilities, especially critical given that record streams (files) will not be present after the
blocknode cutover.
User Story
As an EVM developer, I want my Smart Contract tests to run as fast as possible so that I can
iterate quickly without waiting for transaction confirmations.
Current State vs Target
| Environment | Test Time | Ratio vs Hardhat |
|---|---|---|
| Hardhat local | ~1.18s | 1x (baseline) |
| Local Node (turbo) | ~6.7s | ~6x |
| Solo (current) | ~17.5s | ~15x |
| Solo (target) | ≤ 6.7s | ≤ 6x |
EVM Test Breakdown (Solo Current: 40s total)
| Operation | Solo Current | Target |
|---|---|---|
| Contract deployment | 13.7s | Match Local Node |
| Token minting | 8.2s | Match Local Node |
| Transfer operations | 16.0s | Match Local Node |
| Balance verification | 0.3s | ✅ Acceptable |
Definition of Done
- EVM test suite execution time ≤ Local Node with turbo mode enabled
- Turbo mode equivalent implemented for Solo (accelerated block times)
- Contract deployment, minting, and transfer operations optimized
- Works with blocknode setup (no dependency on record streams/files)
- Benchmark comparison test added to CI
- Developer documentation for enabling turbo/fast mode
Business Justification
- Developer experience: 15x slower than Hardhat is unacceptable for daily development
- Migration path: EVM developers won't adopt Solo if tests are painfully slow
- Competitive positioning: Must match Local Node to be a viable replacement
- Blocknode readiness: Must work without record streams post-cutover
- March MVP deadline: P0 requirement for one-shot EVM deployment
Technical Considerations
- Local Node turbo mode accelerates block production and reduces confirmation times
- Need equivalent mechanism in Solo/Kubernetes environment
- Blocknode architecture changes may require different optimization strategies
- Mirror node propagation delays are a significant contributor
Related
- Parent Epic: Create solo one-shot evm deploy/destroy command #3299 (Create solo one-shot evm deploy/destroy command)
- Benchmark data: [hedera-evm-lab warm cluster](https://github.com/torfinnolsenpersonal/hedera-evm-
lab/blob/main/docs/12-warm-cluster-benchmark-results.md)
Reactions are currently unavailable