[R] R on 64-bit Windows

[Prof Brian Ripley]

On Wed, 15 Oct 2008, Göran Broström wrote:

> In the R for Windows FAQ (2.7.2) we have under 2.2:
> "There is no specific version for x64 Windows, but the standard 32-bit
> version works well enough."
> Does that include the handling of huge data sets?

Yes, but 'huge' is not the largest size in the Huber/Wegman 
classification.  Also, few people with enormous datasets use Windows.

> What I have read in other places indicates that the answer is "No".  If 
> so, will there be a 64-bit version availble soon? When I websearch for 
> 64-bit MinGW it seems to exist out there.

Please do read all the FAQ: this is covered in Q8.1.  You are welcome 
to try building a 64-bit version using the work that has already been 
done, and you are welcome to pay for someone else to do so (and I 
understand that such projects are under way).  If I wanted to do this I 
would buy the PGI compilers and expect the port to be relatively simple.
I've wasted far too much time on mingw64: several times I have built a 
complete version of R and it just crashed or failed even to start.

> -- 
> Göran Broström

-- 
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 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595