[R] Machine accurancy.
ripley at stats.ox.ac.uk
Sun Jun 9 13:43:54 CEST 2002
On Sun, 9 Jun 2002, Kenneth Cabrera wrote:
> Hi, R-users and developers!
> I want to know if it is possible to modify some numeric precision on a
> windows based machine.
> I mean, .Machine() function give me some limits of precision based on my
> machine and OS, but
> I would like to know if I can modify those characteristics from R.
> May I change some C++ library by other and recompile R?
No. And R does not use any C++. Most of the characteristics are of the
ix86 chip and IEEE754 arithmetic. You might be able to use
64-bit-mantissa doubles, if you could find a runtime (`libc') that
> Is it possible to use some specific processor characteristics?
> I'm using R in a W2K O.S, and with an Intel PIII, machine.
> Are those characteristics R better on a LINUX O.S. machine on the same
> PIII processor?
No, they are the same.
> Is it better to use R in the LINUX platform than in W2K with regards
> this machine precision?
Not really. The Linux runtime might be a little more accurate.
Why do you want to do this? R's precision is quite sufficient for almost
all scientific purposes, if used correctly.
> Thank you very much for your help
> double.eps 2.220446e-16
> double.neg.eps 1.110223e-16
> double.xmin 2.225074e-308
> double.xmax 1.797693e+308
> double.base 2
> double.digits 53
> double.rounding 5
> double.guard 0
> double.ulp.digits -52
> double.neg.ulp.digits -53
> double.exponent 11
> double.min.exp -1022
> double.max.exp 1024
> integer.max 2147483647
> sizeof.long 4
> sizeof.longlong 8
> sizeof.longdouble 12
> 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
Brian D. Ripley, ripley at stats.ox.ac.uk
Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel: +44 1865 272861 (self)
1 South Parks Road, +44 1865 272860 (secr)
Oxford OX1 3TG, UK Fax: +44 1865 272595
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