[Biomod-commits] BIOMOD - Problem with memory allocation

Wilfried Thuiller wilfried.thuiller at ujf-grenoble.fr
Fri Nov 20 15:52:41 CET 2009

Dear Anna,
Running a model on almost three millions of points is somehow  
problematic. R is not able to manage nicely so large datasets.

Few comments:

- do you really need to model 2700000 points? Is the data mostly  
consisting of 0? If yes, I would strongly suggest to use the pseudo- 
absence function and repeat the models on sub-fractions of the data.  
This is a much better practice.
- If yes, you have to run the models on all the points because it  
mostly consist of 1 and 0. Split the data in several pieces and then  
run the models sequentially and these sub-data. Then, make the  
projections on all points from each usb-data calibrated models. The  
good things is that you'll have a nice uncertainty analysis at the  
same time.
- Are you sure to run the models on 20 environmental data? I can't  
believe you have 20 uncorrelated variables. I would suggest to run a  
correlation analysis and remove correlated variables which do not  
bring anything excepted noise and memory allocations.

Hope it helps,

Le 20 nov. 09 à 15:07, Anna Cord a écrit :

> Dear all,
> I am trying to run BIOMOD over a large area in South-East Asia...  
> and of
> course I am facing a serious problem with the memory allocation (| 
> error
> message: cannot allocate vector of length...)|
> ||
> |I already tried the suggestions on the R website as wells as  
> Wilfrieds
> suggestion:|
> ||
> "If you use Tinn-R and open R from Tinn-R with the convenient icone,
> there is a very simple way to correct the problem. Go in "Options",  
> then
> "Applications". Then in the R icone, type the following
> --max-mem--size=5000000000000 in the parameters text window."
> The total number of cells is 2525533 and I am working on 1 Species  
> for a
> set of 20 environmental variables.
> Thus, the object size of my input data (X,Y, environmental predictors,
> species presence/absence) is 272759632 bytes - obtained with  
> object.size().
> Do you think by any chance it is possible to run the models on this  
> PC:
> Intel(R) Core(TM)2 Duo CPU E6850 @3.00GHz, 2.99 GHz, 3.25 GB RAM
> Would be very grateful for any help,
> Anna
> University of Würzburg
> Germany
> _______________________________________________
> Biomod-commits mailing list
> Biomod-commits at lists.r-forge.r-project.org
> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/biomod-commits

Dr. Wilfried Thuiller
Laboratoire d'Ecologie Alpine, UMR CNRS 5553
Université Joseph Fourier
BP53, 38041 Grenoble cedex 9, France
tel: +33 (0)4 76 63 54 53
fax: +33 (0)4 76 51 42 79

Email: wilfried.thuiller at ujf-grenoble.fr
Home page: http://www.will.chez-alice.fr
Website: http://www-leca.ujf-grenoble.fr/equipes/tde.htm

FP6 European MACIS project: http://www.macis-project.net
FP6 European EcoChange project: http://www.ecochange-project.eu

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.r-forge.r-project.org/pipermail/biomod-commits/attachments/20091120/c888ca90/attachment.htm 

More information about the Biomod-commits mailing list