inla.exe crashing

[Paul Lantos]

I've been getting INLA crashes despite the fact that these models had previously run successfully. Here is the end of the verbose output and the error message. This is an aspatial version of a model that had worked fine.

formula_2a <- y ~ 0 + b0 + neg + pos +
  f(id, model= "iid", hyper = list(theta = list(prior="pc.prec", param=c(0.5,0.01))))

model_2a <- inla(formula_2a,
                data=inla.stack.data(stk.full.a),
                family="binomial",
                control.predictor=list(A =inla.stack.A(stk.full.a), compute=TRUE),
                control.compute = list(dic = TRUE, waic=TRUE, config=TRUE, cpo=TRUE),
                verbose=TRUE)

	... Parameters for improved approximations
		Number of points evaluate:	 9
		Step length to compute derivatives numerically:	 0.000100002
		Stencil to compute derivatives numerically:	 5
		Cutoff value to construct local neigborhood:	 0.0001
	Log calculations:	 On
	Log calculated marginal for the hyperparameters:	 On
	Integration strategy:	 Automatic (GRID for dim(theta)=1 and 2 and otherwise CCD)
		f0 (CCD only):	 1.100000
		dz (GRID only):	 0.750000
		Adjust weights (GRID only):	 On
		Difference in log-density limit (GRID only):	 4.00
Error in inla.inlaprogram.has.crashed() : 
  The inla-program exited with an error. Unless you interupted it yourself, please rerun with verbose=TRUE and check the output carefully.
  If this does not help, please contact the developers at <he...@r-inla.org>.
In addition: Warning message:
running command '"C:/Program Files/R/R-3.3.2/library/INLA/bin/windows/64bit/inla.exe"  -b -s -v  "C:/Users/pl39/AppData/Local/Temp/RtmpY7T8Os/file1f909cf4433/Model.ini"' had status 253 

[Håvard Rue]

I don't know what this is. there hasn't been any changes in related
part of the code that I'm aware of. I guess you have tried to reboot
the machine. I can rerun it here to check...?

H

>   If this does not help, please contact the developers at <help@r-inl

> a.org>.
> In addition: Warning message:
> running command '"C:/Program Files/R/R-
> 3.3.2/library/INLA/bin/windows/64bit/inla.exe"  -b -s
> -v  "C:/Users/pl39/AppData/Local/Temp/RtmpY7T8Os/file1f909cf4433/Mode
> l.ini"' had status 253 
>

--
Håvard Rue
Department of Mathematical Sciences
Norwegian University of Science and Technology
N-7491 Trondheim, Norway
Voice: +47-7359-3533 URL : http://www.math.ntnu.no/~hrue
Mobile: +47-9260-0021 Email: havar...@math.ntnu.no

R-INLA: www.r-inla.org

[Paul Lantos]

Thanks,

Yeah, I've rebooted. The dataset has confidential data unfortunately so can't share it. Will try to reinstall inla and see if that helps.

Paul

[Håvard Rue]

On Fri, 2017-05-05 at 12:36 -0400, Paul Lantos wrote:
> Thanks,
>
> Yeah, I've rebooted. The dataset has confidential data unfortunately
> so can't share it. Will try to reinstall inla and see if that helps.
>

can you create a fake dataset which also crash?

[Paul Lantos]

I can give it a try -

 The model that is crashing is identical to one that worked, though I've taken out the predictor matrix and SPDE object (to get average / aspatial output). I ran four consecutive INLA models -- model 1 is spatial / unadjusted, model 2 is spatial / adjusted for certain covariates, model 1a is an unadjusted aspatial model, and model 2a is adjusted / aspatia. They all have the same dataset, same mesh, same prediction grid, same SPDE, same priors, etc. But the first three models ran fine (though slowly) and the last one is what crashed.

Paul