[R] Running R on a grid engine

[Huntsinger, Reid]

Does the grid process faster because there is more than one machine or
because each machine (or at least one machine) is faster? If the former,
you're asking about splitting an R process into several processes (to take
advantage of the grid), and you will want to look at the various R packages
to facilitate that: Snow, Rmpi, Rpvm, TaskPR and perhaps others (browse
cran.rproject.org).

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 Adaikalavan Ramasamy
Sent: Thursday, October 28, 2004 2:41 PM
To: S Peri
Cc: R-help
Subject: Re: [R] Running R on a grid engine


See comments below.

On Thu, 2004-10-28 at 18:49, S Peri wrote:
> Dear Group, 
>  I am using DEAL package for modeling signal
> transduction nets.  This process is deal slow on a
> SunFire server with 8 gigs ram. 
> 
> we have a grid that can process much faster that one
> individual server.  
> 
> However, to start the process in Grid, I have to give
> a command or submit a batch process. 
> 
> Is there any way, I can run R in bach process. 
> 
> I tried the following:
> 
> R CMD | library(deal) | data <-
> data.frame(read.table('file1',header=TRUE,
> row.names=1))
> 

Something like this works in *NIX

  echo " print(mean(rnorm(10))) " | R --no-save

HOWEVER you will run into nightmares soon trying to backslash all the
quotes and other special characters. And try to recall a long sequence
of commands you typed in a few days ago ...


Better to put all your codes into a file, say script.R and do

 R --no-save < script.R > log_script.R  &

See help("BATCH"). 


> Here I do not know know :
> 
> 1. How can I point my data files to a function. 

Two ways :
1) Hard code the path inside the script.R
2) Take advantage of commandArgs(). 

Example. If your script.R contains 

data.path <- as.character( commandArgs()[3] )
print(data.path)
load (data.path)  # or read.delim or whatever
....

Then you can pass the path to test.rda via command line

 R --no-save < script.R /home/speri/data/test.rda > log_script.R  &

One thing to keep in mind is to pass the absolute paths and not relative
paths (e.g. ../../data/test.rda). Using relative path may not always
work.


> 2. creating a function (other than in R environment)

Huh ? You can put all your functions into a file, say functions.R, and
you can source("/path/to/functions.R") in your script.R/

> 
> Could any one help me. 
> 
> Thank you in advance.
> 
> 
> Cheers
> Peri.
> 
> ______________________________________________
> 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
>

______________________________________________
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