R-alpha: R-0.49: 2 problems

Peter Dalgaard BSA p.dalgaard@kubism.ku.dk
29 Apr 1997 14:35:36 +0200


Ross Ihaka <ihaka@stat.auckland.ac.nz> writes:

> 
> Kurt Hornik writes:
>  > I found the following 2 problems (no debugging, sorry).
>  > 
>  > * qt(0.975, 3) never returns.
>  > 
>  > Can someone please check that?  Happens for me on Debian GNU/Linux/ix86
>  > compiled with CFLAGS="-O2 -g".
> 
> I am having a look at the computations of p-values and quantiles which
> depend on the incomplete beta function at  present. ....

The original problem seems to be one of loss of accuracy with
instruction reordering. Doubling the value of "acu" cures it (I ran it
under the debugger and the test value was stuck at about 1.5e-32,
being compared to acu=1e-23). This is probably a generic problem with
these last-bit-accurate routines.


-- 
   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-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
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-