Henrik,
thanks for the comments.
> Just so we are one the same page, deleting .Rcache/ directory does not
> make R consume more memory, but rather less.
To clarify, deleting of the cached files was not done due to the
memory issues per se, but because the processing sometimes was
stalling app
Hi.
On Thu, Sep 3, 2009 at 11:37 PM, mbaudis wrote:
>
> Dear Henrik,
>
> FYI, I just ran a complete analysis of 83 SNP6 arrays, based on the
> script at
>
> http://groups.google.com/group/progenetix/web/aroma-affymetrix-processing-visualization
>
> This time, no memory allocation problems. So if
Dear Henrik,
FYI, I just ran a complete analysis of 83 SNP6 arrays, based on the
script at
http://groups.google.com/group/progenetix/web/aroma-affymetrix-processing-visualization
This time, no memory allocation problems. So if this comes up (as it
has before): 64 bit (system + R) seems to help
On Thu, Sep 3, 2009 at 9:17 AM, mbaudis wrote:
>
>> Did you try my 3rd suggestion?
>
> Sorry; I had read this as a future option.
>
> Works like a charm ...
Ok, that's good know. FYI for the future, the CRAN repository will
always be several days delayed relative to the braju.com repository,
whi
> Did you try my 3rd suggestion?
Sorry; I had read this as a future option.
Works like a charm ...
Thanks / Tack!
Michael.
--~--~-~--~~~---~--~~
When reporting problems on aroma.affymetrix, make sure 1) to run the latest
version of the package, 2) to report the
Hi.
On Thu, Sep 3, 2009 at 8:30 AM, mbaudis wrote:
>
> Henrik,
>
> the 2.9.0 is the latest 64 bit compatible binary.
>
> Reason for insisting on 64 bit is the large number of memory problems
> during probe level/fragment length normalization, which require
> frequent process restarts (with R rest
Henrik,
the 2.9.0 is the latest 64 bit compatible binary.
Reason for insisting on 64 bit is the large number of memory problems
during probe level/fragment length normalization, which require
frequent process restarts (with R restart after Rcache deletion),
sometimes already after ca. 20 x 6.0 a
On Thu, Sep 3, 2009 at 4:29 AM, mbaudis wrote:
>
> Dear Henrik,
>
> I am again having problems (new 64bit R 2.9.0 from
> http://r.research.att.com/).
Argh, say what you want about OSX, but it does certainly not make
repository maintenance easy ;)
BTW, R2.9.2 is out now, so 2.9.0 may contain sev
Dear Henrik,
I am again having problems (new 64bit R 2.9.0 from http://r.research.att.com/).
This time, the version of aroma.core seems to be not in place. Thanks
for having a look at it ;-)
Michael.
> hbLite("aroma.core")
Using repository: http://www.braju.com/R/repos
Identified packages to b
Henrik,
since it is fixed, and on a now running production machine, I don't
want to explore too much today. However, I recall that:
1. (definitely) aroma.core was not installed
2. (possibly) R.filesets were not installed, since (I think that) they
downloaded & installed after issuing hbLite("R.f
Thanks Mark,
sorry I wasn't clear. The hbLite() command will not install missing
CRAN/BioC packages, which is why you would expect an error if you
remove required CRAN packages such as digest. Those hbLite() commands
were only intended for troubleshooting Michael's setup.
Michael, also, after
Dear Henrik and Mark,
digest didn't seem to be the problem; I got the same error after
installing it through BioC. However,
> hbLite("aroma.core");
> library("aroma.core");
did load the package. But running
hbInstall("aroma.affymetrix")
complained again about missing aroma.core:
Tags:
Updati
Hey Henrik.
If I run those commands, they work for me, but only because I already have
digest, etc. installed. In order for it to give an error on my R 2.9.1,
I've removed a few packages (e.g. digest) from my default install
directory (/Users/mrobinson/Library/R/2.9/library/), so that it require
Hi,
this seems to be an OSX issue and I cannot reproduce it myself. From
the error message:
it looks like it there are some parsing errors when parsing version
numbers in the PACKAGES file on the repository server, but not sure.
See if this is only a problem with the aroma.affymetrix packa
14 matches
Mail list logo