Performance with context
HashNET V2’s integration suite includes throughput workloads using four or seven validators and parallel transaction submission. The local workload uses signed native transfers, with a default measured interval of several minutes. A same-host test is not a measurement of a geographically distributed mainnet.
What a useful report includes
Section titled “What a useful report includes”- The exact implementation commit and active protocol configuration.
- Validator count, hardware, network topology and latency.
- Transaction type, payload size, gas use and sender count.
- Warm-up, measured interval and load-generation strategy.
- Finalized transactions per second, separately from submitted or accepted throughput.
- Finality latency distribution, errors, rejections and unfinished transactions.
- CPU, memory, storage and network utilization.
Compare like with like
Section titled “Compare like with like”Keep the workload and environment fixed when comparing changes. Repeat runs and investigate variance. A saturated generator can be the bottleneck; confirm the nodes, not the client, are limiting the result.
Do not extrapolate a native-transfer workload to every smart contract. Contract complexity, storage writes and network conditions all affect throughput.
Public results
Section titled “Public results”No verified public-mainnet TPS figure is published on this page. Ask for a reproducible report with the conditions above when evaluating performance. Contact the Tolar team for benchmark and source access.