R help - Mar 2006 - malloc: vm_allocate(size=381886464) failed (error code=3)

Hi all,

I am having memory allocation problem with my R 2.2.1 for Mac OS. The
following is the error message that I get. I do not get this message if I
break down the large dataset in to sub datasets. I think breaking up the
dataset is not a sustainable solution in the long run. The data that I am
analysing is essentially big, and it would be reasonable to do the analyis
on the whole dataset without even considering to partition it. So I was
really wondering if you could give me a clue about how to handle this
problem.

model6 <-lm(logintens~ factor (slide) + factor(ind) + factor(dye) + 
factor(id)*factor(l6) + factor(rep)-1, data=sample_data2)

*** malloc: vm_allocate(size=381886464) failed (error code=3)
*** malloc[387]: error: Can't allocate region
*** malloc: vm_allocate(size=381886464) failed (error code=3)
*** malloc[387]: error: Can't allocate region
*** malloc: vm_allocate(size=381886464) failed (error code=3)
*** malloc[387]: error: Can't allocate region
*** malloc: vm_allocate(size=381886464) failed (error code=3)
*** malloc[387]: error: Can't allocate region
*** malloc: vm_allocate(size=381886464) failed (error code=3)
*** malloc[387]: error: Can't allocate region


Thanks for your help in advance

Regards

Mahdi

-- 
-----------------------------------
Mahdi Osman (PhD)
E-mail: m_osm at gmx.net
-----------------------------------

Echte DSL-Flatrate dauerhaft f?r 0,- Euro*!

On 3/8/06 6:58 AM, "Mahdi Osman" <m_osm at gmx.net> wrote:
> Hi all,
> 
> I am having memory allocation problem with my R 2.2.1 for Mac OS. The
> following is the error message that I get. I do not get this message if I
> break down the large dataset in to sub datasets. I think breaking up the
> dataset is not a sustainable solution in the long run. The data that I am
> analysing is essentially big, and it would be reasonable to do the analyis
> on the whole dataset without even considering to partition it. So I was
> really wondering if you could give me a clue about how to handle this
> problem.
> 
> model6 <-lm(logintens~ factor (slide) + factor(ind) + factor(dye) +
> factor(id)*factor(l6) + factor(rep)-1, data=sample_data2)
> 
> *** malloc: vm_allocate(size=381886464) failed (error code=3)
> *** malloc[387]: error: Can't allocate region
> *** malloc: vm_allocate(size=381886464) failed (error code=3)
> *** malloc[387]: error: Can't allocate region
> *** malloc: vm_allocate(size=381886464) failed (error code=3)
> *** malloc[387]: error: Can't allocate region
> *** malloc: vm_allocate(size=381886464) failed (error code=3)
> *** malloc[387]: error: Can't allocate region
> *** malloc: vm_allocate(size=381886464) failed (error code=3)
> *** malloc[387]: error: Can't allocate region
Mahdi,

How large are the 'id' and 'l6' factors--you are looking for all
interaction
terms?  

As an aside, these look like microarray data.  If they are, is there a
reason that you couldn't use one of the bioconductor packages to do the
analyses (look at limma and maanova, in particular)?  Also, is there a
reason not to do a gene-by-gene analysis?

Sean