[R] Installing R on RedHat7.0

[Thomas Lumley]

On Fri, 3 Nov 2000, Prof Brian D Ripley wrote:

> On Fri, 3 Nov 2000, Tom Wainwright wrote:
> 
> > Thanks!  Building from source worked, and the demos all run.  Several
> > others report that R runs smoothly on RedHat7.
> > 
> > However, not all the tests in $R_HOME/tests succeed.  When I do "make
> > test-Examples" in that directory,  it fails executing "base-Ex.R" in the
> > "eigen" section, with the error:
> > 
> > 	Error: abs(sm %*% V - V %*% diag(lam)) < 60 * Meps is not TRUE
> > 
> > This appears to be a numerical precision problem.  Checking, Meps (i.e.
> > .Machine.double.eps) is 1.084202e-19.  Is this a reasonable value for a
> > Pentium III?
> 
> No. .Machine$double.eps should be about 2e-16.  An i686 does not have
> 19 digits of precision.  Here's what gcc-2.95.2 gives on a PIII under
> RH6.2:
<snip> 


> The likely story is that that beta compiler is even farther from
> IEEE-compliant than is usual for gcc on Linux.  Some of us are
> too nervous about it to even try an installation.

As the value is 2^-63 it may be that the machine precision calculation is
being done entirely internally to the chip, using the 80-bit extended
double format, which has this level of precision.

This is certainly wrong and new; I don't know enough about the machine
precision calculations to know if it is really a compiler bug or whether
it is something the compiler was previously allowed to do but didn't.

	-thomas

Thomas Lumley
Assistant Professor, Biostatistics
University of Washington, Seattle

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._