• Add support for rgamma and rnbinom, using new dust
  • Avoid stack overflow when generating very long expressions
  • More efficient GPU code generation (#75)
  • data_t declarations can use real_t rather than double (#57)
  • GPU support (#66)
  • Start implementing dust gpu support by generating interleaved update function, optionally (#37)
  • Support for dust compare functions (#51)
  • Implement modulo (%%) operator, which uses std::fmod
  • Correct age-structured model in package vignette
  • Separate internal data into mutable and non-mutable components (#45)
  • Support for vectors of floats (rather than doubles) (#6)
  • Add support for config(include) <- file.cpp for including custom C++ code to extend the library support (#39)
  • Add support for odin-like coef(), which returns information about user parameters (#32)