My apologies, call dblepr works precisely as it should my test function neglected to declare the crucial variable double.....
url: www.econ.uiuc.edu/~roger Roger Koenker email rkoen...@uiuc.edu Department of Economics vox: 217-333-4558 University of Illinois fax: 217-244-6678 Urbana, IL 61801 On Jun 16, 2010, at 4:24 PM, Roger Koenker wrote: > I'm trying to understand why 64B R produces an insane answer/plot > to the first > > example(crq) > > in my quantreg package, whereas 32B R produces a sane answer > with the same data. This involves some fortran and in my usual > primitive fashion I've been trying to debug with fortran printing > using the tried and true call dblepr. However, in my test function > call dblepr prints values like > z > [1] 1.644287e-313 > > even though the function returns the sane, correct values. Is this > a known issue, and if so is there an alternative strategy for > printing from fortran in 64b R? > > TIA for any enlightenment. > Roger > > PS. In the hope that someone will tell me that all would be well > if I only followed directions and upgraded R, I'll sheepishly admit: > >> sessionInfo() > R version 2.11.0 Under development (unstable) (2010-02-18 r51149) > x86_64-apple-darwin9.8.0 > > > PPS. The fact that I can run 32B R on the same machine by simply > typing R --arch i386 is one of the great advances of modern science > in my opinion, or at least I thought so, until I discovered this discrepancy > in what was being computed by the two versions. > > > url: www.econ.uiuc.edu/~roger Roger Koenker > email rkoen...@uiuc.edu Department of Economics > vox: 217-333-4558 University of Illinois > fax: 217-244-6678 Urbana, IL 61801 > > _______________________________________________ > R-SIG-Mac mailing list > R-SIG-Mac@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/r-sig-mac _______________________________________________ R-SIG-Mac mailing list R-SIG-Mac@stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/r-sig-mac
