gmwm
gmwm copied to clipboard
Parallelization and RNG
Parallelization
To speed up the bootstraps, we can parallelize different portions of the code. In order to do that, we need to enable stored seeds. This is a more concrete fix of #80.
My thoughts:
- Launch a sitmo R package with linkto headers. [x] see: coatless/sitmo
- Add
seed =
to generating functions. - Remove calls to R::r* in favor of std::gen* or use boost (ziggurat will be faster for normal)