[Rd] cannot increase memory size to 4Gb (PR#11087)
ripley at stats.ox.ac.uk
ripley at stats.ox.ac.uk
Fri Apr 4 19:45:41 CEST 2008
This is (sort of) correct, and as documented. What you cannot increase to
4Gb is the internal *limit* on the allocated memory size, not the 'memory
size' in any sense.
The rw-FAQ in R 2.6.2 patched says
'It can be set to any amount between 32Mb and 4095Mb.'
as does
http://cran.r-project.org/bin/windows/base/rw-FAQ.html#There-seems-to-be-a-limit-on-the-memory-it-uses_0021
The maximum size of memory for a process under Vista64 is 4Gb (4096Mb),
*but* R cannot allocate all of it (not least because the R executable and
DLL need to fit into it).
Version 2.4.1 said (correctly)
'It can be set to any amount between 32Mb and 3Gb',
because versions of Windows current at the time were limited to a 3Gb
address space (and so the real limit was a little less than 3Gb).
On Fri, 4 Apr 2008, n.henriquez at ion.ucl.ac.uk wrote:
> Full_Name: Nick Henriquez
> Version: 2.6.2; 2.6.1 and 2.4.1
> OS: Vista business 64
> Submission from: (NULL) (144.82.49.16)
>
>
> I try to increase memory size (by copy-pasting from FAQ)
> --max-mem-size=1Gb at the end of the Target field (after any final
> double quote, and separated by a space) and changing the 1 to a 4.
>
> After double-clicking the shortcut (for modified 2.4.1 OR 2.6.1 OR
> 2.6.2) I get --max-mem-size=4G: too large and ignored.
>
> I have Vista business64 with 8 Gb of RAM. BioC2.1 and 2.2 complain about
> a lack of memory.
Maybe, but that is not an R-bugs issue.
>
> Suggestions welcome but I think I am "alone" so far.
R-help is available to you to ask questions about things you do not
understand -- you will get a much more sympathetic response if you only
use R-bugs for things you 'know for sure' are not doing what they should.
--
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-devel
mailing list