[BioC] problem with arrayQualityMetrics: cannot allocate vector

[Wolfgang Huber]

Hi Jarek

I could not reproduce this problem on two machines available to me 
(x86_64-unknown-linux-gnu (64-bit), with 8 GB RAM, and 
x86_64-apple-darwin10.4.0/x86_64 (64-bit), with 4 GB RAM). You also 
mentioned (outside this list) that the report was produced without 
errors on a different machine of yours (Ubuntu 10.10 64bit 8GB RAM 3Gz i7).

So it appears that this problem is quite specific to your system, 
i386-apple-darwin9.8.0/i386 (32-bit). Note that the amount of RAM that 
you have there is no less than what I have. Perhaps you're hitting a 2 
GB limit (on a 32-bit OS) for your R process (?).

It seems there are three options, which are all on your end:
- update the R on your Mac to 64 bit
- use a bigger machine
- subsample the object 'x', e.g.
    sx = x[ sample(nrow(x), 5e4), ]
   which should give you an almost as useful report. Also, you probably 
want to set 'do.logtransform = TRUE'.


(Btw, object.size(x) is 271675096 bytes ~ 260 MB, which is only a 
fraction of the available RAM. Nevertheless, I watched the R process 
grow to 2.5-3 GB while munching through 'arrayQualityMetrics()', which 
is a consequence of that function creating a couple of copies of the 
data - minimisation of memory use has not been a design goal so far.)

	Best wishes
	Wolfgang

Jarek Bryk scripsit 08/11/10 14:07:
> Hello,
> the problem is as in the subject, and here are the details:
>
>> class(RG)
> [1] "RGList"
> attr(,"package")
> [1] "limma"
>> library(arrayQualityMetrics)
>> library(convert)
>> x=as(RG,"NChannelSet")
>> class(x)
> [1] "NChannelSet"
> attr(,"package")
> [1] "Biobase"
>
>> arrayQualityMetrics(x,outdir="qc")
> The report will be written in directory 'qc'.
> Error: cannot allocate vector of size 74.3 Mb
> R(2581,0xa048a500) malloc: *** mmap(size=3899392) failed (error code=12)
> *** error: can't allocate region
> *** set a breakpoint in malloc_error_break to debug
> R(2581,0xa048a500) malloc: *** mmap(size=7794688) failed (error code=12)
> *** error: can't allocate region
> *** set a breakpoint in malloc_error_break to debug
> R(2581,0xa048a500) malloc: *** mmap(size=3899392) failed (error code=12)
> *** error: can't allocate region
> *** set a breakpoint in malloc_error_break to debug
> R(2581,0xa048a500) malloc: *** mmap(size=7794688) failed (error code=12)
> *** error: can't allocate region
> *** set a breakpoint in malloc_error_break to debug
> R(2581,0xa048a500) malloc: *** mmap(size=77922304) failed (error code=12)
> *** error: can't allocate region
> *** set a breakpoint in malloc_error_break to debug
> R(2581,0xa048a500) malloc: *** mmap(size=77922304) failed (error code=12)
> *** error: can't allocate region
> *** set a breakpoint in malloc_error_break to debug
> R(2581,0xa048a500) malloc: *** mmap(size=77922304) failed (error code=12)
> *** error: can't allocate region
> *** set a breakpoint in malloc_error_break to debug
>
>> sessionInfo()
> R version 2.12.0 (2010-10-15)
> Platform: i386-apple-darwin9.8.0/i386 (32-bit)
>
> locale:
> [1] en_US.UTF-8/en_US.UTF-8/C/C/en_US.UTF-8/en_US.UTF-8
>
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
>
> other attached packages:
> [1] convert_1.26.0            marray_1.28.0             limma_3.6.6               arrayQualityMetrics_2.6.0
> [5] affyPLM_1.26.0            preprocessCore_1.12.0     gcrma_2.22.0              affy_1.28.0
> [9] Biobase_2.10.0
>
> loaded via a namespace (and not attached):
>   [1] affyio_1.18.0        annotate_1.28.0      AnnotationDbi_1.12.0 beadarray_2.0.1      Biostrings_2.18.0
>   [6] DBI_0.2-5            genefilter_1.32.0    grid_2.12.0          hwriter_1.2          IRanges_1.8.2
> [11] lattice_0.19-13      latticeExtra_0.6-14  RColorBrewer_1.0-2   RSQLite_0.9-2        simpleaffy_2.26.0
> [16] splines_2.12.0       stats4_2.12.0        survival_2.36-1      vsn_3.18.0           xtable_1.5-6
>
>
>
> RG is a result of importing 10 samples into limma from agilent scanner. Each sample is a 1M probe mouse custom CGH array. Is it a memory/performance problem? I have 4Gb of RAM + 2.4GHz Core2 Duo processor. I would be very grateful for any clues as to what is wrong and how to fix it...
>
> Regards,
> jarek
>


-- 


Wolfgang Huber
EMBL
http://www.embl.de/research/units/genome_biology/huber