Problem of vector allocation in GSIF package

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Problem of vector allocation in GSIF package

Manuel Spínola
Dear list members,

I am fitting a model with the GSIF package, but I ran into a problem of
vector allocation.  Is there any way to solve this problem?  See code and
error message below.

I am using:
R 3.6.1
GSIF 0.5-5
Mac with 16 GB of RAM

> rk_rf_ac <- fit.gstatModel(variables_todos_sp["ac"], ac_formulaString_correlacion, covar_finales_sp, method = "quantregForest")Fitting a Quantile Regression Forest model...Shapiro-Wilk normality test and Anderson-Darling normality test report probability of < .05 indicating lack of normal distribution for residualsFitting a 2D variogram...Saving an object of class 'gstatModel'...> rk_rf_ac_pred <- GSIF::predict(rk_rf_ac, covar_finales_sp, predict.method = "KED")Error: cannot allocate vector of size 19.4 Gb

Thank you very much,

Manuel

--
*Manuel Spínola, Ph.D.*
Instituto Internacional en Conservación y Manejo de Vida Silvestre
Universidad Nacional
Apartado 1350-3000
Heredia
COSTA RICA
[hidden email] <[hidden email]>
[hidden email]
Teléfono: (506) 8706 - 4662
Personal website: Lobito de río <https://sites.google.com/site/lobitoderio/>
Institutional website: ICOMVIS <http://www.icomvis.una.ac.cr/>

        [[alternative HTML version deleted]]

_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
Reply | Threaded
Open this post in threaded view
|

Re: Problem of vector allocation in GSIF package

Tomislav Hengl-5

You have two options:

1. Run predictions using tiling
(https://github.com/Envirometrix/BigSpatialDataR#dem-analysis-using-tiling-and-parallelization)

2. Buy more RAM.

I suggest using option 1 since option 2 can propagate to infinity.

PS: I am working on a new package
(https://github.com/Envirometrix/landmap/) that should give more
flexibility to users and maybe even incorporate tiling of large objects
by default.


On 9/13/19 5:38 PM, Manuel Spínola wrote:

> Dear list members,
>
> I am fitting a model with the GSIF package, but I ran into a problem of
> vector allocation.  Is there any way to solve this problem?  See code and
> error message below.
>
> I am using:
> R 3.6.1
> GSIF 0.5-5
> Mac with 16 GB of RAM
>
>> rk_rf_ac <- fit.gstatModel(variables_todos_sp["ac"], ac_formulaString_correlacion, covar_finales_sp, method = "quantregForest")Fitting a Quantile Regression Forest model...Shapiro-Wilk normality test and Anderson-Darling normality test report probability of < .05 indicating lack of normal distribution for residualsFitting a 2D variogram...Saving an object of class 'gstatModel'...> rk_rf_ac_pred <- GSIF::predict(rk_rf_ac, covar_finales_sp, predict.method = "KED")Error: cannot allocate vector of size 19.4 Gb
>
> Thank you very much,
>
> Manuel
>

_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
Reply | Threaded
Open this post in threaded view
|

Re: Problem of vector allocation in GSIF package

Manuel Spínola
Thank you very much Tom.

I will like to try landmap.

Best,

Manuel

El vie., 13 sept. 2019 a las 9:45, Tomislav Hengl (<[hidden email]>)
escribió:

>
> You have two options:
>
> 1. Run predictions using tiling
> (
> https://github.com/Envirometrix/BigSpatialDataR#dem-analysis-using-tiling-and-parallelization
> )
>
> 2. Buy more RAM.
>
> I suggest using option 1 since option 2 can propagate to infinity.
>
> PS: I am working on a new package
> (https://github.com/Envirometrix/landmap/) that should give more
> flexibility to users and maybe even incorporate tiling of large objects
> by default.
>
>
> On 9/13/19 5:38 PM, Manuel Spínola wrote:
> > Dear list members,
> >
> > I am fitting a model with the GSIF package, but I ran into a problem of
> > vector allocation.  Is there any way to solve this problem?  See code and
> > error message below.
> >
> > I am using:
> > R 3.6.1
> > GSIF 0.5-5
> > Mac with 16 GB of RAM
> >
> >> rk_rf_ac <- fit.gstatModel(variables_todos_sp["ac"],
> ac_formulaString_correlacion, covar_finales_sp, method =
> "quantregForest")Fitting a Quantile Regression Forest model...Shapiro-Wilk
> normality test and Anderson-Darling normality test report probability of <
> .05 indicating lack of normal distribution for residualsFitting a 2D
> variogram...Saving an object of class 'gstatModel'...> rk_rf_ac_pred <-
> GSIF::predict(rk_rf_ac, covar_finales_sp, predict.method = "KED")Error:
> cannot allocate vector of size 19.4 Gb
> >
> > Thank you very much,
> >
> > Manuel
> >
>


--
*Manuel Spínola, Ph.D.*
Instituto Internacional en Conservación y Manejo de Vida Silvestre
Universidad Nacional
Apartado 1350-3000
Heredia
COSTA RICA
[hidden email] <[hidden email]>
[hidden email]
Teléfono: (506) 8706 - 4662
Personal website: Lobito de río <https://sites.google.com/site/lobitoderio/>
Institutional website: ICOMVIS <http://www.icomvis.una.ac.cr/>

        [[alternative HTML version deleted]]

_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
Reply | Threaded
Open this post in threaded view
|

Re: Problem of vector allocation in GSIF package

Manuel Spínola
In reply to this post by Tomislav Hengl-5
I wanted to say, "I would like to try landmap".

El vie., 13 sept. 2019 a las 9:45, Tomislav Hengl (<[hidden email]>)
escribió:

>
> You have two options:
>
> 1. Run predictions using tiling
> (
> https://github.com/Envirometrix/BigSpatialDataR#dem-analysis-using-tiling-and-parallelization
> )
>
> 2. Buy more RAM.
>
> I suggest using option 1 since option 2 can propagate to infinity.
>
> PS: I am working on a new package
> (https://github.com/Envirometrix/landmap/) that should give more
> flexibility to users and maybe even incorporate tiling of large objects
> by default.
>
>
> On 9/13/19 5:38 PM, Manuel Spínola wrote:
> > Dear list members,
> >
> > I am fitting a model with the GSIF package, but I ran into a problem of
> > vector allocation.  Is there any way to solve this problem?  See code and
> > error message below.
> >
> > I am using:
> > R 3.6.1
> > GSIF 0.5-5
> > Mac with 16 GB of RAM
> >
> >> rk_rf_ac <- fit.gstatModel(variables_todos_sp["ac"],
> ac_formulaString_correlacion, covar_finales_sp, method =
> "quantregForest")Fitting a Quantile Regression Forest model...Shapiro-Wilk
> normality test and Anderson-Darling normality test report probability of <
> .05 indicating lack of normal distribution for residualsFitting a 2D
> variogram...Saving an object of class 'gstatModel'...> rk_rf_ac_pred <-
> GSIF::predict(rk_rf_ac, covar_finales_sp, predict.method = "KED")Error:
> cannot allocate vector of size 19.4 Gb
> >
> > Thank you very much,
> >
> > Manuel
> >
>


--
*Manuel Spínola, Ph.D.*
Instituto Internacional en Conservación y Manejo de Vida Silvestre
Universidad Nacional
Apartado 1350-3000
Heredia
COSTA RICA
[hidden email] <[hidden email]>
[hidden email]
Teléfono: (506) 8706 - 4662
Personal website: Lobito de río <https://sites.google.com/site/lobitoderio/>
Institutional website: ICOMVIS <http://www.icomvis.una.ac.cr/>

        [[alternative HTML version deleted]]

_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo