R help - May 2003 - lme: cannot allocate vector

Hi,
I am trying to fit a linear mixed effects model (lme) with a variance 
covariance matrix that has 148248^2 entries. I get the error message: 
error: cannot allocate vector of size 2243241 kb

However, I am running this on a sun sparc (solaris 2.8) with 20GB of ram, 
using R 1.7
Is this a hard limit I am reaching or does this have to do with how sun 
segments memory?

I can probably rewrite the model to be a bit more sensible with how 
resources are used,
but I thought i could brute force it since we have the power.

Thanks
Nicholas
	[[alternate HTML version deleted]]

On Mon, 26 May 2003 LEWIN_KOH_NICHOLAS_JOHN at lilly.com wrote:
> Hi,
> I am trying to fit a linear mixed effects model (lme) with a variance 
> covariance matrix that has 148248^2 entries. I get the error message: 
> error: cannot allocate vector of size 2243241 kb
> 
> However, I am running this on a sun sparc (solaris 2.8) with 20GB of ram, 
> using R 1.7
> Is this a hard limit I am reaching or does this have to do with how sun 
> segments memory?
Did you build a 64-bit version of R (see the R-admin manual)?  If not, 
you are limited to 4Gb total, and that's the likely explanation.

-- 
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