Quantum Computing & Engineering · Practical tool · English

Make a quantum pilot comparison reproducible

Start with a problem another person can reproduce. Run a small exact classical baseline, then record the complete timing and quality context of an experiment you have actually performed.

Free to explore. Worksheet inputs stay in this browser tab; no account or paid API is required.

Conceptual title illustration for Quantum Computing & Engineering
Conceptual title image. Any real observation data is identified separately below.
No quantum service is called.

The baseline below is a real browser calculation on a six-node fixture graph. Quantum fields are for your own measurements. This page does not execute a quantum circuit, fabricate a hardware result or establish quantum advantage.

Run an exact classical Max-Cut baseline

Divide the six nodes into two sets to maximise the sum of edge weights crossing between them. With one node fixed to remove complementary duplicates, the exact solver evaluates 32 assignments. Green edges cross the returned partition; node colours identify its two sets.

Download graph and timing protocol
No baseline run yet.

Browser timing varies with device, load and timer resolution. For this tiny graph it is not a meaningful production speed benchmark. The exhaustive solution is useful as a correctness reference, while the fixed graph and method make the result reproducible.

Record a complete workflow comparison

Enter measured seconds for the same problem instance and an agreed success criterion. Include repeated-run budgets and objective quality in the downloaded record or your experiment log. Encoding, queue time and readout cannot be silently dropped from the quantum side.

No comparison recorded.

Make the experiment inspectable

Minimum comparison record
DimensionRecord explicitlyWhy it matters
ProblemGraph, weights, objective direction and constraints.A different instance is a different comparison.
QualityCut objective, success criterion, seeds, shots and repeat count.Faster output is not useful if it misses the agreed target.
ResourcesClassical hardware; simulator or QPU; compilation; queue; execution; readout.The full workflow may dominate a small circuit.
ReproducibilitySoftware versions, configuration, raw results and analysis.Another reviewer needs to understand what was actually run.

A useful result may be that this problem does not justify quantum experimentation. That is an engineering decision, not a failed marketing claim. For background, read the original QAOA paper, while keeping its algorithmic proposal separate from any result entered here.

Take the next step with context

Keep the question, assumptions and evidence together. These tools help organise a review; an exported record does not approve a design or certify a system.