Release of FLUKA 4-3.4

Dear FLUKA users,

It is our pleasure to announce that a new minor release of the FLUKA code, version 4-3.4, has been issued and can be downloaded from the Official CERN FLUKA website:

See the release notes at the bottom of this message for a detailed account of all novelties.

For consistency, please update to the latest version of Flair, available from

With our kindest regards,

The FLUKA developer team at CERN and Collaborating Institutes


=== Fluka-4.3.4 ===

The code is now aligned with DPMJET v19.3.6, fixing a rare compilation/linking issue.


- Point-wise low-energy neutron interactions:

    * In previous versions, a too relaxed search tolerance of +-10 K precluded a correct selection of the requested S(a,b,T) temperature for H bound in ortho and para H2 and D2 environments, for which database temperatures are 1 K apart. As a result, the lowest database temperature was inadvertently loaded. As of this version, thermal scattering laws are loaded for the temperature closest to the one requested by the user, as intended.

    * Fixed a bug in the S(a,b,T) thermal scattering law treatment, whereby the total interaction cross section was correctly computed, but at the moment of deciding the interaction channel (elastic, inelastic, or capture), unbound cross sections were invoked instead of bound S(a,b,T) when requested. On the other hand, the actual sampling of the final state was performed according to the bound S(a,b,T) cross sections - simply with an altered rate. The effect of this issue was noticeable up to the meV range.

    * Fixed a rare bug whereby an elastic recoil was left with slightly negative kinetic energy due to numeric roundoff.

- Fixed two rare crashes in rQMD (kindly reported by BNL colleagues) which were traced back to numerical roundoff issues.

- In the crystal channeling module, several redundant matrix operations were optimized, leading to an enhanced simulation efficiency.

- Fixed a bug, whereby in case of multiple SR photon emission along a single charged-particle step, the position of all but the first SR photon was erroneously sampled, altering the radius of curvature along the foreseen step.
This affected only extreme situations, e.g. TeV e- in magnetic fields of O(10) T.

- Resolved a rare crash due to a charged particle produced at an energy higher (yet kinematically still valid) than the upper energy abscissa for dE/dx tabulations, thus fixing the crash reported in this forum thread.

- Reviewed the patch for ionization fluctuation underflow originally reported in v4-3.2.