0

Rcpp and RcppArmadillo are useful for R package development. As for RcppArmadillo, it seems superlu is the only method to utilize the sparse property for very large matrices. However, it is not clear how to let an R package with superlu work in Windows.

Can anyone give me an example in which R package using superlu works well in Windows? I think I can check the source code to figure it out.

Thanks, Wenjian

Wenjian Bi
  • 11
  • 3
  • That is indeed a good question. As a first proxy, can you maybe work out how to a) compile SuperLU locally with Rtools and b) put value 'by hand' into `src/Makevars.win` to use the location? – Dirk Eddelbuettel Mar 01 '21 at 13:31
  • Thank you Dirk. I understand the whole process. But, to be honest, I am not familiar with compilation. Previously, I also met similar problems and I usually go find an R package and copy the thirdParty folder with source code and then update `src/Makevars.win`. That is exactly why I asked this question. I was thinking there should be R package using superlu but was a little bit shocked that I did not find any one through google search. I think superlu is very useful and the only reason might be superlu cannot be easily used in Windows. – Wenjian Bi Mar 01 '21 at 15:09

0 Answers0