This is the v2.0.0 release of Galacticus, following semantic versioning. Use it with the corresponding v2.0.0 releases of:
Changes between v1.0.0 and v2.0.0
New features
- For progenitor mass function analyses allow single-parent-mass and variable root-variance
- Add output analysis classes for quiescent fractions and star forming main sequences
- Add support for CLASS transfer functions
- Add output of the transfer function in the power spectrum task
- Add a new nodePropertyExtractor class to extract the optical depth in Lyman-alpha through the ICM
- Add a new analysis class to compute the CGM/ICM X-ray luminosity-temperature relation
- Replace "host history" and "age statistics" node components with meta-properties
Bug fixes
- Avoid attempts to access null hotHalo component
- Ensure scalar property data is deallocated for rank1 properties (and vice-versa)
- Allow build to succeed with meta-properties and GALACTICUS_ACTIVE_COMPONENTS
- Correct logic in analysis plotting script that determines y-axis range
- Internal I/O should be performed inside an OpenMP critical section
- Escape pound sign in URL
- "boolean" property operator should never return -1
- Allow the "normal" property operator to correctly handle cases with infinite extent
- Covariance model for volume functions and cross-correlation term for mean functions
- Makes zero integral kernel no longer an error
- Pass root-variance (instead of variance) to cosmologicalMassVariance class mass function
- Re-use tabulated excursion set solutions
- Mark nodes with too large scale radii as unsolvable
- Adjust tolerance in Ludlow2016 concentration test
- Add missing quarter-mode mass function to BBKS transfer function class
- Fix array bounds error
- Avoid retabulation of delta_crit(t) during merger tree building
- Pass "node" object to critical overdensity functions to allow environment to be accessed
- Evaluate age-weighted integrals only during inactive property solving
- Ensure doc build scripts respect the BUILDPATH environment variable
- Avoid floating point exception due to rounding errors
- Handle cases where the negative binomial distribution success probability is 1
- Remove deprecated XML attributes
- Allow "description" as element in "method" element
- Correct documentation of the black hole-bulge mass analysis class
- Avoid division by zero error in Local Group mass function analysis if no host halos were found
- Do not output attributes associated with target dataset if no such dataset exists
- Avoid attempting to read non-existant "logLikelihood" attributes
- Improve the help message for failed broad-band luminosity integration
Performance improvements
- Use the complementary error function instead of (1-erf(x)) in the Farahi and midpoint excursion set solvers
- Replace the use of approximate error function with the intrinsic error function in Fortran, which provides higher accuracy at similar speed
- Replace the use of error function from GSL with the intrinsic Fortran function.
- Optimize likelihood calculation for "modelLikelihoodGalaxyPopulation" class
- Improve HDF5 dataset chunking