R help - Aug 2002 - AW: Out of memory

> "Liaw, Andy" wrote:
> > 
> > I have run into similar problems occasionally on Windoze 
> NT.  For whatever
> > reason, R refuses to do anything, including q("no").  
> 
> So you got a prompt and were able to type it in? Strange! Never seen
> something like that! 
> 
> 
> > The only way out was
> > to kill it via Task Manager.  Not sure how to make it reproducible.
> > Unfortunately Uwe's comments are not exactly helpful, 
> though mostly true.
> 
> a) What was wrong?
OK, the memory.limit(...) command was unreasonable as your comment pointed out.
The version hint may also prove helpful.

But you did not comment all the reasonable commands I typed in (Andy added on
another one) without getting a reasonable response from R. Why not? Thus it
seemed to me that you would expect and accept that kind of behaviour of R: E.g.
loading some data into memory which exceed the recent memory size will result in
refusing any further command (more precise: getting a prompt and an invariable
response to any typed command). This is not exactly helpful...
> 
> b) They were helpful, since, obviously, the usage of 
> memory.limit() was
> not reasonable.
> 
> c) Some more un-helpful comments on memory usage:
> One of my collegeaus asked me how to compute all possible permutations
> of the integers 1, 2, ..., 20 and save them to a file. Well, 
> I told him,
> in principle he could use ... But he should think about the memory
> usage, since these are 20! permutations * 20 integers to store. If he
> could use just one byte per integer, approx. 50000000000 
> GigaByte would
> be required. ;-)
Maybe helpful for him. Even though not exactly the focus of my contribution. 
> 
> Best,
> Uwe
> 
Nikolaus Hansen
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

On Mon, 19 Aug 2002, Nikolaus Hansen wrote:
> But you did not comment all the reasonable commands I typed in
(Andy added on another one) without getting a reasonable response from R.
Why not? Thus it seemed to me that you would expect and accept that
kind of behaviour of R: E.g. loading some data into memory which exceed
the recent memory size will result in refusing any further command
(more precise: getting a prompt and an invariable response to any
typed command). This is not exactly helpful...

Would you prefer the process to be terminated?  You used up all the
resources allocated to R.  None could be freed by garbage collection.
What was R supposed to do?  There was no memory left to do anything else
in.

I think this circumstance is very rare (normally you get enough memory
back to do something).  I would prefer it to allowing R to take over all
the VM on a computer, when you probably need to reboot. Those who have not
seen this happen on a departmental server (say) should consider themselves
fortunate.  (It was open to you to set any limit you like if you prefer
this behaviour.)

-- 
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 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._