Choose and install nuclear data
ACTINV software and evaluated nuclear data have separate versions. The embedded catalog identifies fixed release artifacts; installing software does not silently replace your calculation’s data.
Standard neutron setup
actinv data fetch
actinv data verify
actinv data list
The default bundle is tendl-2025-neutron, the full 2,850-target neutron library. Catalog 1.1.0 installs it under actinv-data/v1.1.0/ with its index, ENDF/B-VIII.0 primary decay data, and JEFF-3.3 fallback decay data.
Downloads are checked by size and SHA-256 before installation. Correct existing files are reused. An incorrect existing file is left in place unless you explicitly request replacement with --force.
Choose a bundle
| Calculation | Activation bundle | Matching covariance bundle |
|---|---|---|
| Full neutron corpus | tendl-2025-neutron | tendl-2025-neutron-covariance |
| Derived patched neutron subset | tendl-2025-patched-neutron | tendl-2025-patched-neutron-covariance |
| Proton | tendl-2025-proton | None in the shipped catalog |
| Deuteron | tendl-2025-deuteron | None in the shipped catalog |
| Alpha | tendl-2025-alpha | None in the shipped catalog |
For example:
actinv data fetch tendl-2025-neutron-covariance
actinv data verify tendl-2025-neutron-covariance
A covariance sidecar must match the exact activation library and index. Both full and patched neutron corpora have matching sidecars; mixing them is rejected. Add the corresponding uncertainty section to your problem to request propagation. Downloading a sidecar alone does not enable uncertainty.
The patched subset corrects a bounded upstream defect and excludes many targets. It is not a drop-in accuracy upgrade for every material or spectrum. Read Known data limitations before choosing between the neutron bundles.
Use another location
actinv data fetch --output /data/actinv
actinv data verify --output /data/actinv
actinv new problem.json --data-dir /data/actinv
--data-dir writes absolute paths to the generated problem. To keep symbolic catalog references, set ACTINV_DATA_DIR to your installation root when using the default actinv new output. It names the root above v1.1.0, not the version folder itself.
actinv data fetch and verify use --output for their destination; pass it explicitly when installing or checking a custom location.
Offline installation
Run actinv data manifest to print the embedded catalog. Obtain the listed archives and extracted artifacts through your usual transfer process, retain their names and versioned paths, then run actinv data verify on the destination machine. See the source and extraction record for provider details.
Prepared calculation cache
The CLI creates a verified prepared cache when it first uses a library and spectrum. Later runs reuse those compact files. The public iron example creates about 282 MiB of prepared files; other libraries and spectra differ.
By default, ACTINV uses the platform’s cache location. Set ACTINV_CACHE_DIR to an absolute path to choose another disk. This cache is disposable; removing it causes data preparation to run again. Keep the original nuclear-data inputs, problem, and result for the calculation record.
A corrupt or incompatible final cache artifact produces an error instead of being silently trusted. Troubleshooting explains how to regenerate it. Desktop workers use a private cache per calculation.
Terms and custom data
The installed ACTINV-DATA-NOTICE.md records data attribution and transformations. ACTINV’s MIT/Apache-2.0 software licence does not replace a dataset’s licence or source terms.
Advanced users can build activation, covariance, shielding, and damage tables from evaluated files. See Advanced workflows and the specification reference.