Dear FLUKA users,

It is our great pleasure to announce that a new major release of the FLUKA code, version 4-5.0, has just been issued and can be downloaded from the Official CERN FLUKA website: https://fluka.cern.

Please see the release notes at the bottom of this message for a detailed account of all new features, among which we wish to highlight new technical capabilities, including the long-awaited support for unstructured mesh geometries (externally generated from CAD/CAE files in TetGen, Abaqus, or ANSYS cdb formats) and an experimental automatic biasing algorithm, as well as rich physics improvements featuring an improved gamma de-excitation treatment for thermal neutron capture.

To comfortably profit from the new features, please update to the latest Flair v3.4.X, available in the Official Flair website.   

Note that a webinar showcasing the highlights of this major release is planned to take place in April. A further communication will follow in the course of March to announce the modalities under which registration to this event will be arranged.

We take this opportunity to remind you that our next beginner course will be kindly hosted by the European Commission's Joint Research Centre (JRC) in Geel, Belgium, from 12 to 16 May 2025. We can accommodate only a few more participants, so do not delay if you wish to attend. You can find all relevant information on the Indico page of the course. If you have any question regarding our training programme, please contact fluka.course@cern.ch.

With kind regards,

The FLUKA developer team at CERN and Collaborating Institutes

=== Fluka-4.5.0 ===

This major release includes powerful technical improvements, featuring support for unstructured-mesh geometries (externally generated from CAD/CAE geometries) and a beta version of an automatic importance biasing algorithm, as well as a series of relevant physics improvements, among which enhanced gamma cascades from thermal-neutron capture. See the list below for a detailed account of all physics and technical improvements, as well as bug fixes.

PHYSICS IMPROVEMENTS:

- Gamma emission from thermal-neutron capture events is now initiated from ENSDF data blocks for 168 isotopes, duly taking into account the competition between the emission to discrete levels and to the continuum.

- Furthermore, in the gamma de-excitation module, a crude fallback estimate of the energies of the first and second nuclear levels has been deprecated in favor of actual nuclear level data when available, thereby removing spurious gamma lines.

- Point-wise interactions of low-energy neutrons for which the evaluated nuclear data prescribe a residual nucleus break-up are now explicitly treated by dedicated call to the FLUKA Fermi break-up module. In our present point-wise neutronics databases, this concerns non-elastic channels for 6Li, 7Li, 10B, 12C, and 14N. NB: for residual 8Be (in the ground state or in any excited level) or 12C (in excited levels only), a call to Fermi break-up is issued regardless of break-up flags. This improvement leads to an observable suppression of spurious gammas in favor of particle emission; residual nuclei production from the aforementioned channels is thereby also improved.

TECHNICAL IMPROVEMENTS:

• A new capability has been added to support unstructured meshes (UM) extracted from CAD/CAE files in TetGen, Abaqus, or ANSYS cdb format via the UMESH card. This memory-efficient, high-performance CPU implementation is based on Flair’s GeoViewer module, enabling users to include multiple UMs in a single input while allowing transformations and CSG operations on UMs. UM regions can be treated on equal footing to standard FLUKA CSG regions. For details on usage, refer to the dedicated manual page, as well as the Unstructured Mesh guide in our documentation section. To fully benefit from this feature, please ensure you have the latest version of Flair (3.4.*) installed.
• An automatic importance biasing algorithm has been introduced as an experimental feature, greatly facilitating the creation and use of importance biasing schemes at the minimal expense of issuing the newly included card AUTOIMBS. Please refer to the manual page of the latter for its usage, as well as for relevant notes and a disclaimer elucidating its current experimental status.
• A new command-line utility (fluresnucle_formatter.py) has been added in the bin folder for the conversion of tab.lis results files from RESNUCLE scorings into the following formats: txt, tex, json, csv, xlsx (Excel).
• In past FLUKA versions, the absence of a RANDOMIZe card would imply that every cycle in a run would start with the same seed. This behavior has now been forbidden: absence of a RANDOMIZe card is now equivalent to having a RANDOMIZe card with default WHATs, thereby ensuring that subsequent cycles continue on the random-number sequence prescribed in the first cycle and avoiding the pointless repetition of identical cycles observed on several occasions in this forum.
• The magnetic horn has been included among the analytical magnetic fields of cards MGNCREATe and MGNFIELD. See the manual page of the MGNCREATe card for the mplemented analytical expression of this field.
• In the CRYSTAL card, a reference momentum is now offered as an option for crystal property initialization at a momentum other than that of the BEAM card (the default value).
• Mac OSX compilation and linking:
  - The build issue reported in this forum thread has been tentatively fixed by the explicit addition
  of the `-ld_classic` linker flag, an option which will be dropped in the near future.
  - Optimize-sibling calls have been disabled at compilation for the function performing point-wise neutron inelastic interactions in order to avoid a miscompilation.
• The peak RAM usage since the start of the run is now logged in the .out file.
• The codebase of FLUKA's Fermi break-up and gamma de-excitation modules has been migrated to a modern C++ framework, in view of the ongoing evolution towards the next FLUKA generation.

BUG FIXES:

Proton interactions:

• In earlier FLUKA versions, nuclear elastic scattering of protons was inactive below 1.1 MeV for all nuclei. However, the Coulomb barrier for target nuclei with Z<12 is lower than 1.1 MeV (it may be of even a few 100 keV). As of FLUKA v4-5.0, nuclear elastic scattering of protons is now also active below 1.1 MeV on nuclei with lower Coulomb barrier.
• Fixed a bug kindly reported by BNL colleagues, wherein a p+p inelastic interaction failed to complete, being performed in inverse kinematics with the inverse boost leading to an infinitesimally negative kinetic energy.

Point-wise neutronics:

• Overcame an issue whereby photons emitted during fission-fragment de-excitation carried an incorrect statistical weight. 
• Fixed a bug whereby pre-scission neutrons were always emitted from the fission fragments instead of the compound nucleus, thereby slightly altering the kinematics.
• Fixed the low-energy neutron counter issue pointed out in this forum post.         
• Implemented perfomance enhancements to sampling from Legendre expansions.
• Fixed a rare numerical-precision error leading to a division by zero in spurious zero-length tabular intervals.

Group-wise neutronics:

• Fixed a bug whereby in the group-wise treatment of 14N, the expected proton from 14N(n,p) was not explicitly generated.

Deuteron interactions:

• Adapted the d+T reaction cross-section such that, for Ed<3.75 MeV, only (d,n) from IAEA parametrization is sampled, thereby removing a small fraction of spurious (d,ng) events. A slight excess of low-energy (d,n) events has likewise been overcome.
• Fixed a bug whereby, in a narrow slice of phase space, the compound formed after stripping to the continuum of one of the nucleons eventually yielded photons with incorrect statistical weight.

Decay database:

• Fixed a bug whereby the beta decay lines of 11C were taken as those of 11Be, and the two internal conversion electron lines of 137La were taken as the first two of 137Cs.
• Fixed a bug whereby 11Be lines were flagged as an attempted beta+ decay in lieu of the now correct beta- behavior.
• In the crystal channeling module, a spurious division by hbar*c has been removed in the wavevector-transfer evaluated in the channeling form factor routine, whereby slightly excessive channeling efficiencies at 100s of GeV and above have been overcome.

Scoring:

• Fixed a bug whereby USRYIELD cards with empty intervals for the second quantity led to a crash, as reported in this forum thread. The code now gracefully stops with a message pointing to the problematic USRYIELD card, also in case the interval for the first quantity is empty after all initialization logics.
• Gaussian energy broadening: if the parameters a, b, and c are such that FHWM = a + b * sqrt(dE + c (dE)**2) is negative in some interval of dE, no Gaussian broadening is applied (as opposed to crashing as in earlier FLUKA versions) and a warning message is written to the output file the first time such a situation is encountered for every DETECT scoring. Addresses the inconvenience reported in this forum thread.

Miscellaneous:

• In very rare cases, during the synchrotron radiation emission, a (now lifted) excessively strict check on the momentum conservation was causing unwarranted aborts.
• Overcame a numerical round-off issue in the matrix coefficients for rotations involving angles below the microradian.