R-beta: R has underflows on sparc-redhat-linux

Peter Dalgaard BSA p.dalgaard at biostat.ku.dk
Tue Jul 28 00:11:15 CEST 1998


Jeff Johnson <jbj at redhat.com> writes:

> 
> Hi --
> 
> I'm trying to package R for inclusion on some upcoming Red Hat
> Powertools distribution. I have started with the 0.61 version
> packaged by Martyn Plummer <plummer at iarc.fr> (Thanks Martyn!)
> and have upgraded to 0.62-2. I have succeeded in running R demo's
> on  i386-redhat-linux-gnu, and now I am also trying to get the
> alpha and sparc versions working.

Interesting. If you have anything that looks like a deadline for this,
we'd like to know. This kind of thing might get us to modify our
schedules to get a release out so that we can avoid known-and-fixed
bugs to find their way onto widespread CD's (and pester the mailing
lists for years after...).

> On sparc-redhat-linux, I'm seeing a SIGFPE with the following
> stacktrace:
[snip: R dies computing y * y with:]
> (gdb) p y
> $1 = 7.4583407312002067e-155
> (gdb) 
> 
> Although I understand what the cause of the problem is, I'm not sure
> what the correct thing to do to enable underflow traps on sparc-redhat-linux.

Eh? (That sentence doesn't quite parse...) Generally, R relies on IEEE
FPU behaviour - as is default on i386 architecture Linux. I.e. SIGFPE
should not be generated upon arithmetic underflow - or any FPU
condition for that matter.

If you look in the R-devel archives (on CRAN), there was a note from
Goran Brostrom on June 4 saying that he had R running on RedHat
Sparc/4.2. 

> In addition, I notice that the configure for R does not try and
> compile with (on alpha) -mieee nor link with "-lieee -lm".  Is that
> OK?

Probably not. Getting R to compile on Alpha (Linux and Digital Unix)
was hashed out only recently, and the proper compile flags may not
have found  their way into configure.in yet. Some notes on R-devel and
R-help around the end of June relate to this.

If you get these matters sorted out, we'd be only too happy to fix
configure accordingly.

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help 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-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._



More information about the R-help mailing list