Performance

HPEC exists to serve workloads that need high throughput and low, predictable latency while still settling on-chain. Performance is tracked against explicit targets and exercised with Altius's benchmarking tooling.

Figures are targets, not published results

The figures below are design targets used for internal validation. Treat them as targets to confirm against your own benchmark runs before citing them externally.

Targets

Metric	Target
Sustained throughput	10,000+ TPS
Execution latency	< 1 ms
Confirmation latency	< 1 ms
Settlement to Altius L1	< 60 ms
Scaling	Horizontal — scales with hardware, not validators

Latencies are per stage: execution and confirmation happen off-chain in sub-millisecond time, while settlement batches verified results to Altius L1. Deployments can be tuned to your workload, or hardware-optimized to exceed these numbers.

Methodology

Benchmarks drive synthetic workloads across configurable batch sizes and load levels and record throughput and latency distributions. Runs are executed against a representative deployment so results reflect the full execute-then-settle path, not just isolated execution.

Note

Exact target values, percentiles, and degraded-mode thresholds depend on your deployment's configuration and hardware.