On 25 October 2024 at 15:42, Denis Haine wrote: | I'm totally puzzled on why I've got these results. How can I get the same | output with Rcpp + regression as with the R loop +/- regression?
I think you continue to gloss over the fact there "is a contract": we call C++ code from R, Rcpp helps, Rcpp aligns the RNG state and restores it, this all works well. You violate (somehow, not entirely sure how/why) this contract via your pattern of calling _back into R_ from your C++ code. And I think that part is the source of your problem. I think I suggested last time that you work it out with smaller and smaller test cases. That still seems like a good idea to me. Good luck, Dirk -- dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org _______________________________________________ Rcpp-devel mailing list Rcpp-devel@lists.r-forge.r-project.org https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel
