Skip to content

Memory question on R

3 messages · Dr L. Y Hin, Brian Ripley, roger bos

Original
Dr L. Y Hin
# 
Dear all,
Apology for posting this amateurish question.
I am running R version 2.1.0 on Windows XP for a simulation exercise.
Seemingly, I've encountered memory allocation problem during the 
recursive procedure.
I've looked at the help section called 
?Memory() and ?memory.limit(), and ?memory.size but the comamands
seem to refer to Unix version of R.
I wound be very grateful if anyone could kindly advise me on the method 
to set the memory allocation so that R workspace can harness the maximum
resources available on the hardware (RAM and virtual memory), with an understanding
that my machine may not have sufficient resources to run other applications.
In addition, does R version 2.1.0 makes use of the hyperthreading feature that may 
speed up the application?

Thanking you in advance
Lin
Brian Ripley
#
On Fri, 20 May 2005, Dr L. Y Hin wrote:

            

      
      
      
    

        
If you looked in a Windows version of R they do not (and they should not 
have the ()).

Please do read the posting guide and (as we ask there) show us exactly 
what the messages you got were.  This question seems more appropriate to 
R-help than R-devel.

            

      
      
      
    

        
We do ask for no HTML mail.

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      roger bos
    

    

      
      #
    

  


  

      
Lin,

I don't know about hyperthreading, but on my 4GB XP machine I can read
in very large data files to the extend that Windows Task Manager shows
the Rgui is using 2.3 to 2.7 GB.  Further, I can run automated
simulations on this data all weekend with the processor at 99% without
the system or R crashing.  But to do this I had to modify the header
file to make Rgui /LARGEADDRESSAWARE.  There is a FAQ which describes
how to do this.  Let me know if you cannot find it.

However, it may be worth checking your code and seeing if there is
anyway you can avoid recursiveness as this may be not memory
efficient.

This is definitely a R-help questions.  Rd is for bug reports and such.

Thanks,

Roger

        
On 5/20/05, Prof Brian Ripley <ripley@stats.ox.ac.uk> wrote: