[R-SIG-Mac] Does preprocessCore do ppc64?

Loren Engrav engrav at u.washington.edu
Thu Dec 13 23:58:19 CET 2007


Thank you, was helpful, is complicated for this geezer, but interesting

The whole goal is to ReadAffy() on 81 chips which seems to need 64 bit to
avoid "cannot allocate vector...." error (or reduce chips or analyze in
pieces but not yet ready to give up on this)

Other advisors advised

"Install Leopard 10.5.1

Now install either
    R-2.6-branch-leopard-universal.tar.gz
or
    R-devel-leopard-universal.tar.gz
depending on which version of BioC you want to use: the former is for BioC
2.1 (the current release) and the latter for BioC 2.2 (the current devel).

Then start R with --arch=ppc64 and you will be running in 64-bit
mode."

So I installed Leopard and R-2.6-branch and start with R --arch=ppc64

=========
and as you suggested while running Command Line

> version
               _   
platform       powerpc-apple-darwin9.0.0
arch           powerpc
os             darwin9.0.0
system         powerpc, darwin9.0.0
status         Patched
major          2   
minor          6.1 
year           2007
month          12  
day            12  
svn rev        43661
language       R   
version.string R version 2.6.1 Patched (2007-12-12 r43661)

> .Machine$sizeof.pointer == 8
[1] TRUE

> R.version$arch
[1] "powerpc"

> .Platform$r_arch
[1] "ppc64"

============
Then I run R.app with options(arch="ppc64") in .Rprofile

And get

> version
               _   
platform       powerpc-apple-darwin9.1.0
arch           powerpc
os             darwin9.1.0
system         powerpc, darwin9.1.0
status         Patched
major          2   
minor          6.1 
year           2007
month          12  
day            12  
svn rev        43661
language       R   
version.string R version 2.6.1 Patched (2007-12-12 r43661)

> .Machine$sizeof.pointer == 8
[1] FALSE

> R.version$arch
[1] "powerpc"

> .Platform$r_arch
[1] "ppc"
> 

So would then seem at Command Line I have 64 but in R.app I do not

============
Then I try biocLite("preprocessCore") at Command Line and returns

> .Platform$r_arch
[1] "ppc64"
> sessionInfo()
R version 2.6.1 Patched (2007-12-12 r43661)
powerpc-apple-darwin9.0.0

locale:
en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

loaded via a namespace (and not attached):
[1] rcompgen_0.1-17
> source("http://bioconductor.org/biocLite.R")
> biocLite("preprocessCore")

Running biocinstall version 2.1.8 with R version 2.6.1
Your version of R requires version 2.1 of Bioconductor.
trying URL 
'http://bioconductor.org/packages/2.1/bioc/bin/macosx/universal/contrib/2.6/
preprocessCore_1.0.0.tgz'
Content type 'application/x-gzip' length 157468 bytes (153 Kb)
opened URL
==================================================
downloaded 153 Kb


The downloaded packages are in
    
/var/folders/-O/-O2rx2VN2RWJgU+8ZMxMx++++TI/-Tmp-//RtmpHcRo5R/downloaded_pac
kages
> library(preprocessCore)
Error: package 'preprocessCore' is not installed for 'arch=ppc64'
> 

===============
Then I try same with R.app and returns

> .Platform$r_arch
[1] "ppc"
> sessionInfo()
R version 2.6.1 Patched (2007-12-12 r43661)
powerpc-apple-darwin9.1.0

locale:
en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base
> 

> source("http://bioconductor.org/biocLite.R")
> biocLite("preprocessCore")
Running biocinstall version 2.1.8 with R version 2.6.1
Your version of R requires version 2.1 of Bioconductor.
trying URL 
'http://bioconductor.org/packages/2.1/bioc/bin/macosx/universal/contrib/2.6/
preprocessCore_1.0.0.tgz'
Content type 'application/x-gzip' length 157468 bytes (153 Kb)
opened URL
==================================================
downloaded 153 Kb


The downloaded packages are in
    
/var/folders/-O/-O2rx2VN2RWJgU+8ZMxMx++++TI/-Tmp-//Rtmp3VK9Rw/downloaded_pac
kages


> library(preprocessCore)
Error in dyn.load(file, ...) :
  unable to load shared library
'/Library/Frameworks/R.framework/Resources/library/preprocessCore/libs/ppc/p
reprocessCore.so':
  
dlopen(/Library/Frameworks/R.framework/Resources/library/preprocessCore/libs
/ppc/preprocessCore.so, 5): Library not loaded:
/Library/Frameworks/R.framework/Versions/2.6/Resources/lib/libgfortran.2.dyl
ib
  Referenced from: 
/Library/Frameworks/R.framework/Resources/library/preprocessCore/libs/ppc/pr
eprocessCore.so
  Reason: image not found
Error in library(preprocessCore) : .First.lib failed for 'preprocessCore'

=============
So to me seems with your investigative advice
1) I have 64bit at command line
2) I do not have 64bit in R.app
3) preprocessCore is not 64bit (and other packages)

So I am closer to 64bit ReadAffy of 81 files but still not there

With your serious knowledge, how should I get to 64bit ReadAffy

Thank you





> From: Prof Brian Ripley <ripley at stats.ox.ac.uk>
> Date: Thu, 13 Dec 2007 17:00:36 +0000 (GMT)
> To: Loren Engrav <engrav at u.washington.edu>
> Cc: "r-sig-mac at stat.math.ethz.ch" <r-sig-mac at stat.math.ethz.ch>
> Subject: Re: [R-SIG-Mac] Does preprocessCore do ppc64?
> 
> The short answer is that the BioC binary packages are not 64-bit.  But
> then neither is the CRAN build of R 2.6.1 nor the CRAN binary packages.
> 
> And very likely you are running a 32-bit R.
> 
> On Thu, 13 Dec 2007, Loren Engrav wrote:
> 
>> I tried
>> options(arch="ppc64") in Rprofile
>> and seems to have worked,
> 
> I am sure it 'worked' for some value of 'worked', but what do you think it
> did?  You can't change the architecture of a running R process (and you
> need a running R process to use R code ...).  You find out whether the
> process is 64-bit by checking
> 
> .Machine$sizeof.pointer == 8
> 
> and R.version$arch will also indicate the architecture name (x86_64 on my
> Macintel builds).
> 
> The sub-architecture (if any) of the running process is given by
> .Platform$r_arch and the user cannot change it (it was set at configure
> time).  From the output below, that is 'ppc'.  Not only do the
> architectures need to match, but do the sub-architecture settings: this is
> easy to achieve when building everything yourself, but not when taking
> binaries built on different systems by different people at different
> times.
> 
> So, it seems
> 
> - you are running a 32-bit R.
> - you are missing a Fortran library.
> 
> Please do try to deduce from the output what the problem is, not guess and
> never test your guesses.
> 
>> I think R.app is ppc64 in Leopard
> 
> The CRAN build of 2.6.1 is not, and you haven't told us which build you
> are using.
> 
> AFAIK if you want 64-bit, you need to build packages from sources.
> 
> How to set all this up is AFAIK not fully documented, for the very good
> reason that it is *experimental*.  My understanding is that it is intended
> to support 64-bit MacOS binary builds from R 2.7.0 only.
> 
> 
> 
>> But affy will not load as requires preprocessCore
>> 
>> Library(preprocessCore) returns
>> 
>> Error in dyn.load(file, ...) :
>>  unable to load shared library
>> '/Library/Frameworks/R.framework/Resources/library/preprocessCore/libs/ppc/p
>> reprocessCore.so':
>> 
>> dlopen(/Library/Frameworks/R.framework/Resources/library/preprocessCore/libs
>> /ppc/preprocessCore.so, 5): Library not loaded:
>> /Library/Frameworks/R.framework/Versions/2.6/Resources/lib/libgfortran.2.dyl
>> ib
>>  Referenced from:
>> /Library/Frameworks/R.framework/Resources/library/preprocessCore/libs/ppc/pr
>> eprocessCore.so
>>  Reason: image not found
>> Error in library(preprocessCore) : .First.lib failed for 'preprocessCore'
>> 
>> I check libs in preprocessCore and find i386 and ppc, no ppc64
>> 
>> Does this mean preprocessCore as it stands does not do ppc64? Or I am
>> getting the wrong one? Or I must build 64bit? Or?
>> 
>> The original thread is in Bio but this seems like a Mac specific question
>> 
>> Thank you
>> 
>> 
> 
> -- 
> 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



More information about the R-SIG-Mac mailing list