Read your results

actinv run problem.json result.json saves one JSON document with every computed step, calculation diagnostics, and the input certificate. Keep that file alongside the problem and selected data.

Time steps and units

Each entry in steps reports the end of a schedule segment. step is one-based; t_s is cumulative seconds from the start, including irradiation and cooling.

For a sample of known mass in grams, multiply the per-gram activity or heat by that mass to obtain Bq or W. Check that the composition and its basis describe the material you intend.

total_atoms_per_g is an accounting diagnostic across the solver state vector. It includes sinks and, where present, a unit source state. Use the per-nuclide inventory for physical nuclide populations.

Optional results

Photon sources, dose-response quantities, pathways, uncertainty, radiological responses, and damage observables depend on the requested options and required data. An absent field means that output was not emitted; it is not a zero result.

Photon group values represent source strength integrated over each energy group. Contact gamma dose is a semi-infinite-slab screening proxy; use separate transport for geometry-dependent dose. Radiological indices use your selected coefficient table. NRT dpa uses the supplied damage-energy table and displacement energies. Scope and qualification explains these boundaries.

Pathways identify a source and first product with ranked contributions in trace mode. They do not enumerate every intermediate member of a reaction chain.

Check the ledger

The top-level ledger records missing data, composition issues, leakage, pruning, mode selection, numerical diagnostics, and optional response coverage. Inspect it before interpreting a small value as physical absence.

Positive activity without a radiological coefficient and material targets without damage data are reported as incomplete coverage. For features that support it, require_complete turns missing coverage into a calculation error.

The numerical_floor_atoms_per_g field is a CRAM asymptotic scale, not a bound on total numerical error. Below-floor counts and heat bounds help identify small populations that need closer review.

Read uncertainty with its coverage

An uncertainty response contains its nominal value, propagated standard uncertainty, interval, coverage, and a separate CRAM-order comparison. The default channel is MF=33 cross-section covariance. Half-life and independent fission-yield uncertainties can be enabled explicitly. Current master also supports a requested transport-tally statistical flux channel; check release availability.

Incomplete evaluated covariance does not become complete by requesting a confidence level. Incident-flux uncertainty is excluded unless the flux channel is requested. That channel uses supplied groupwise tally errors and a diagonal model; systematic transport-model, geometry, and nuclear-data errors remain excluded. Composition, response-coefficient, and other model uncertainties also remain outside the band. Aggregate activity uncertainty must be propagated as activity.total; summing individual standard uncertainties does not preserve correlations.

You can declare an additional unmodeled relative term directly or through a hash-pinned calibration/evaluation-spread table. That declared contribution supplements the band; it cannot account for a missing reaction channel or establish complete uncertainty coverage. See additional uncertainty reporting.

Check input identity and comparisons

The certificate records the selected inputs and computed hashes. It establishes which files were used; it does not establish their physical suitability.

Before comparing two results, check the material basis, spectrum normalization, schedule, projectile, library, and decay data. Compare at the same physical time. Record missing-data and mode differences alongside numerical differences.