<340 ns Per decision
<600 ps Logical op
2.9M/sec Throughput
On-premise native No cloud runtime

Performance as Proof of Architecture

Performance is not about speed. It is about determinism under load. When infrastructure must remain predictable at scale, latency variance matters more than raw numbers. Every benchmark below is reproducible, methodology-transparent, and measured on a single core with no cloud dependency.

Benchmark Results

Decision Resolution

Resolve (1 evidence)
304 ns
Resolve (10 evidence)
311 ns
Resolve (100 evidence)
329 ns

Constant time complexity — O(1). Median 300 ns independent of evidence count.

Trilean Logic (K3 Kleene)

Single operations
AND
532 ps
OR
563 ps
NOT
312 ps
Batch · 1000 operations
AND ×1000
1.24 µs
OR ×1000
1.24 µs
NOT ×1000
407 ns

Cryptographic Integrity

Content hash (256 B)
230 ns
Content hash (10 KB)
24.5 µs

Reading note: bar lengths are normalized per group; absolute values are shown on the right.

Criterion.rs · Apple M-series · Single core · Release mode · LTO enabled · 300 samples · Variance < 1.4%

Source: cargo bench · Variance < 5% over 100 iterations

Infrastructure Capacity Simulation

100,000
10K1B
Annual decisions processed 36,500,000
Daily single-core utilization 0.000042%
Growth headroom Extreme

Why This Fast

Speed is a consequence of architectural discipline, not an optimization target.

Native Rust

No garbage collector, no VM, no runtime overhead. Compiled to native machine code.

Stack-only hot path

Zero heap allocation on the decision resolution path. Predictable memory, zero GC pauses.

Constant-time logic

Trilean operations are lookup tables, not decision trees. O(1) regardless of input complexity.

No network dependency

The OmegaOS runtime runs on-premise. No API calls, no cold starts, no cloud latency in the critical path.

Reproducible proof

Every number comes from Criterion.rs, statistically rigorous benchmarks.

Start a Pilot →