[Rd] Error compiling R on RS6000 (PR#475)

Douglas Bates bates@stat.wisc.edu
08 Mar 2000 07:39:59 -0600

Thomas J Vogels <tov@ece.cmu.edu> writes:

>   I'll assume Paul meant config.site which contains all the extra
>   compiler and linker flags and those flags are somewhat involved in
>   his environment.  Triggered by Paul's email I tried to
>   compile R-1.0.0.  It compiles cleanly but dies with "Illegal
>   Instruction" upon start.  Sorry, I don't have any further insight
>   into this problem at this time.

Do the compilers need special flags to control IEEE floating point
behavior?  One of the first things that R does on startup is to force
an evaluation of 1.0/0.0 and store the result as a global "Infinity".
On some architecture/compiler combinations this will generate an error
unless you instruct the compiler to insert special code for full IEEE
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch