[R] going past restrictions on number of elements

[Prof Brian Ripley]

On Mon, 1 Aug 2011, Paul Rodriguez wrote:

> Hello R experts,
>
> I'm trying to test R in a shared memory environment in which addressable memory is aggregrated to about 600G.
>
> However, I get an error of 'too many elements' specified when I try creating a 45K x 100K matrix.
>
> I tried running R with a --max-nsize=50000000000 option, but got the same message.
>
> Is there a way to run create such large matrices?

No.  See ?"Memory-limits" (a matrix in R is also a vector).

NB: setting a limit on Ncells (there normally is not one) isn't going 
to help allocation of vectors, is it?
>
> thanks,
> Paul Rodriguez

-- 
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