[BioC] Memory Issue under WinXP x64 (64 bit Windows XP)

[Kasper Daniel Hansen]

Actually in the R for Windows FAQ (which you of course consulted  
before sending your email), it says under 8.1
"Note that building under Windows 64 is not supported as no suitable  
compiler is available"

This means you need to switch OS if you want R to use the many GB of  
memory you have.

/Kasper

On May 18, 2006, at 2:01 PM, Carleton Garrett wrote:

> Hi
>
> I'm currently running Bioconductor version 2.2.0 under Windows XP x64
> with 16 Gb RAM and Virtual Memory upto 100 Gb.
>
> In trying to combine 67 Affy u133a and 67 Affy u133a_2 cel files I am
> able to form the initial affy batches using read.affybatch() but get a
> memory allocation error (below) when I try to combine them with the
> 'combineAffyBatch()' function
>
> Error: cannot allocate vector of size 280,470 Kb
>
> Attempts to increase memory size within 'R' are only partly  
> successful.
> Attempting to increase the memory to 16 Gb (16000 Mb) through using  
> the
> following command
>
> memory.limit(size=16000)
>
> produces the error
>
> Error in memory.size(size) : don't be silly! : your machine has a 4 Gb
> limit [not true according to the OS].
>
> Reducing the number of cel files in the two initial affy batches to 16
> each (but not 33 each) did allow the combined affy batch to be created
> so the combneAffyBatch() function appears to work on this platform.
>
> Two questions:
>
> 1)   Does the binary version 2.2.0 of R for Windows have a built-in
> memory limit of 4 Gb? - and if so, is there a way to circumvent  
> this limit?
>
> 2)   Is there anything peculiar about the combineAffyBatch() function
> that makes it use memory inefficiently?
>
>     I raise the above question because even when the number of cel  
> files
> was reduced to 33 u133a and 33 u133a_2 there was a memory allocation
> error when I ran combineAffyBatch() that was the same as with the 67 +
> 67 cel files (ie, "Error: cannot allocate vector of size 280,470 Kb")
> and yet the memory.limit() was ~4.2 Gb and the memory.size() was only
> ~1.4 Gb in a machine with 16 Gb RAM.
>
> Any suggestions welcomed.
>
> Carl Garrett
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/ 
> gmane.science.biology.informatics.conductor