Benchmark Results: 186M outputs/J

RTX 3060 Ti · Deterministic reruns · 0 NaN / 0 Inf across all documented validation paths

RTX 3060 Ti
Consumer-accessible hardware
All tiers bit-exact
Tier Comparison

Performance by Tier

All tiers measured at D=1000 on RTX 3060 Ti. Determinism verified across repeated runs.

Quick
QUICK
186M
outputs / joule
Latency 5.755 ms
Depth D=1000
Determinism PASS
NaN / Inf 0 / 0
Standard
STANDARD
187M
outputs / joule
Latency 5.809 ms
Depth D=1000
Determinism PASS
NaN / Inf 0 / 0
Premium
PREMIUM
98M
outputs / joule
Latency 11.174 ms
Depth D=1000
Determinism PASS
NaN / Inf 0 / 0
Validation

Validation Results

All checks performed on the same hardware used for benchmark measurement.

RTX 3060 Ti · Full Validation Pass · Documented Run
NaN count 0
Inf count 0
Determinism checks PASS (repeated runs)
Validation path coverage Documented
Hardware RTX 3060 Ti
Interpretation

What the numbers mean

Plain-language definitions of the reported metrics.

Metric · outputs/J

outputs/J is a combined throughput-per-energy metric. It describes how many signal outputs are produced per joule of energy consumed. Higher is more efficient.

Property · Determinism

Determinism means the same input produces the same output on every run. This is verified by executing the pipeline multiple times and confirming bit-exact output match.

Metric · 0 NaN / 0 Inf

0 NaN / 0 Inf means no numerical instability was observed in documented validation paths. Every output value in the measured runs was a valid finite number.

Context

Pilot context

These figures reflect the current implementation stage on the documented hardware. A feasibility sprint applies the same deterministic pipeline to your specific signal or data-processing use case, and produces benchmark evidence for that exact configuration.

Start a Pilot