Re: [R] 0.64.1 alpha floating point exception in make check

Peter Dalgaard BSA (p.dalgaard@biostat.ku.dk)
08 Jun 1999 10:10:22 +0200

To: Dick Repasky <drepasky@bio.indiana.edu>
Subject: Re: [R] 0.64.1 alpha floating point exception in make check
From: Peter Dalgaard BSA <p.dalgaard@biostat.ku.dk>
Date: 08 Jun 1999 10:10:22 +0200
In-Reply-To: Dick Repasky's message of "Mon, 7 Jun 1999 21:05:00 -0500 (EST)"

Dick Repasky <drepasky@bio.indiana.edu> writes:

> > plot(update(rock.ppr, sm.method="gcv", gcvpen=2),
> + main = "update(..., sm.method=\"gcv\", gcvpen=2)")
> Floating point exception
>
> Is this error known?

No...

> Is it associated with floating point exception handling by Linux kernel
> (2.2.5-16 of Red Hat distribution) on alpha hardware?

I suspect so, or hardware FPU configuration. Normally, we rely on
having default IEEE handling (set flag and continue with NaN/Inf
result) which should *never* produce SIGFPE. There is one exception
though: integer division by zero, which is not maskable and produces
an FP (sic!) error.

Perhaps it would be a good idea if you tried running R under gdb and
see where the exception comes from. ("R -d gdb" and then "run
--vanilla < tests/Examples/modreg-Ex.R")

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard@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@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._


This archive was generated by hypermail 1.02.