[R] R memory allocation error - Unix

Liaw, Andy andy_liaw at merck.com
Fri Nov 21 20:27:40 CET 2003


> From: Susan Shortreed [mailto:susanms at stat.washington.edu] 
> 
> I am using ESS on a unix system for my analysis.  My R 
> environment contains a 90118 by 94 dataframe.  I am trying to 
> calculate the mean of a column in this data frame and I am 
> getting the following error:
> 
> Error: can not allocate a vector of size 704 Kb
> 
> I have tried
> options(memory=1000000000000000000)
> and this does not help.
> 
> when I call gc() this is what is returned
> > gc()
>            used (Mb) gc trigger  (Mb)
> Ncells  1178845 31.5    2564037  68.5
> Vcells 11666683 89.1   38686231 295.2
> 
> I tried calling mem.limits(nsize=1000000000000).  Any value 
> for vsize gives an NA error, and when I recall gc() the limit 
> for Vcells is NA. There is more than enough memory available 
> on the Unix machine, when I call top I am using 0.0% of the 
> memory and the other handful of users are using about 10% all 
> together.  I have increased my  user memory limit and that 
> still did not help (I found an email in R-help archives 
> suggesting this).  It seems to me that 704Kb is a rather 
> small size to give an error and it appears to be available on 
> the system.
> 
> Any suggestions?

What exactly is this "unix" system?  This could be important, because I was
recently made aware of a problem on AIX 5, where R was compiled as 32-bit.
By default the R process will only have access to 256 MB of RAM, even though
ulimit -a reports unlimited and the machines has GBs of RAM.  I had to set
an evironment variable before starting R to get it to use up two 4GB of
memory.

This may not be your problem, but you do need to provide more details about
your system.

Andy

 
> Thank you,
> Susan
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list 
> https://www.stat.math.ethz.ch/mailman/listinfo> /r-help
>




More information about the R-help mailing list