R help - Mar 2005 - about memory

here is my system memory:
ronggui at 0[ronggui]$ free
             total       used       free     shared    buffers     cached
Mem:        256728      79440     177288          0       2296      36136
-/+ buffers/cache:      41008     215720
Swap:       481908      60524     421384

and i want to cluster my data using hclust.my data has 3 variables and 10000
cases.but it fails and saying have not enough memory for the vector size.  I
read the help doc and use $R --max-vsize=800M to start the R 2.1.0beta under
debian linux.but it still can not get the solution.so is my pc'memory not
enough to carry this analysis or my mistake on setting the memory?

thank you.

How much memory is free when R fails (e.g., what does "top" show while
trying to run your clustering)?  If there's still a sizeable amount of free
memory you may have to look into the system limits, maximum data segment size in
particular.  Many Linux distros have it set to "unlimited" but default
Debian may not.  If this turns out to be the problem, please do not, _do not_
raise it to "unlimited," but only to enough for R to work.

hth,

jon b



On Wed, 30 Mar 2005 18:36:37 +0800
ronggui <0034058 at fudan.edu.cn> wrote:
> here is my system memory:
> ronggui at 0[ronggui]$ free
>              total       used       free     shared    buffers     cached
> Mem:        256728      79440     177288          0       2296      36136
> -/+ buffers/cache:      41008     215720
> Swap:       481908      60524     421384
> 
> and i want to cluster my data using hclust.my data has 3 variables and
10000 cases.but it fails and saying have not enough memory for the vector size. 
I read the help doc and use $R --max-vsize=800M to start the R 2.1.0beta under
debian linux.but it still can not get the solution.so is my pc'memory not
enough to carry this analysis or my mistake on setting the memory?
> 
> thank you.
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html

hclust creates a distance matrix. In your case it is 10,000 x 10,000. For
various reasons several copies are created, so you probably need at least 

100M x 8 bytes per entry x 3 copies = 2.4 GB

just for the distance matrix. If you don't have that much RAM the
computation will probably take longer than you're willing to wait.

Reid Huntsinger

-----Original Message-----
From: r-help-bounces at stat.math.ethz.ch
[mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of ronggui
Sent: Wednesday, March 30, 2005 5:37 AM
To: r-help at stat.math.ethz.ch
Subject: [R] about memory


here is my system memory:
ronggui at 0[ronggui]$ free
             total       used       free     shared    buffers     cached
Mem:        256728      79440     177288          0       2296      36136
-/+ buffers/cache:      41008     215720
Swap:       481908      60524     421384

and i want to cluster my data using hclust.my data has 3 variables and 10000
cases.but it fails and saying have not enough memory for the vector size.  I
read the help doc and use $R --max-vsize=800M to start the R 2.1.0beta under
debian linux.but it still can not get the solution.so is my pc'memory not
enough to carry this analysis or my mistake on setting the memory?

thank you.

______________________________________________
R-help at stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html