Unfortunately, I can't tell you how to solve this problem. However, I can
confirm that this crash happens also on my machines (R 2.0.1 on Debian Sarge,
AMD Athlon, 750 MHz, 256 MB RAM and R 2.0.0 on SuSE 9.0, Intel P4, 2.6 GHz,
512 MB RAM):
R> aaa(1000)
alloc of 80000 bytes failed
alloc of 40000 bytes failed
Speicherzugriffsfehler
We had a model, which consists of 1000+ linear programs with each about 550
rows and 1100 columns (http://www.unipr.it/arpa/dipseq/EAAE/PR/Parallel/
42b4_henningsen.pdf). On MS-Windows 2000/XP this model always crashed after
some hundreds linear programs, but using Linux on the same machines it
_never_ crashed. I really can't see the reason why our much larger programs
worked, but these small programs let R crash.
I hope that this information may help to solve this problem.
Arne
On Wednesday 30 March 2005 12:16, TEMPL Matthias wrote:> Hello!
>
> I have a curious problem, which I cannot solve.
> With my code I solve thousands of small linear programs with the package
> lpSolve automatically. But R crashes sometimes (~always, but always on
> different linear programs) in a strange way. For illustration, I tried to
> prepare a simple example, which shows the nature of the problem. The
> function aaa (see below) declares some constants (only in this special
> example) and in the end it solves the linear program.
>
> aaa(1)
> [,1] [,2]
> [1,] 0 0
> [2,] 0 0
> [3,] 0 0
> [4,] 0 0
> [5,] 52 0
> [6,] 2 2
> [7,] 0 0
> [8,] 2 0
> [9,] 0 0
> [10,] 0 0
> [11,] 0 0
> [12,] 0 0
> [13,] 54 0
>
> Works fine.
> Now I make the *same* calculation, say 1000 times:
> aaa(1000)
> R (I have tried it with R2.0.1, 2.0.0, 1.9.1, 2.1.0dev) crashes completly -
> without warning and error message under Windows XP, Intel Pentium 3 with
> 256 MB RAM Under Linux SuSe 8.2 R (2.0.1) it crashes again, but in this
> case I get the following message: Calloc of 40004 bytes failed on line 114
> of file lpkit.c
> ...
> Calloc of 80008 bytes failed on line 113 of file lpkit.c
> Error: cannot allocate vector of size 3 Kb
>
> Now I?m completly lost. Solving the linear program one time makes no
> problem. Solving it twice in the same way makes no problem either. Running
> the same calculation, say 1000 times, causes a crash. Why should there be a
> problem with memory?
>
> For any hint, I would be really happy.
> Thank you,
> Matthias
>
>
> ### ---------- function aaa ------------
>
> aaa <- function(amount=1){
> f.obj <- rep(0,33)
>
> w <- c(3,4,5,6,11,13,17,22,25,26,27,28,33)
>
> m <-
> matrix(c(0,0,1,1,1,1,0,0,0,0,-1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
>,
>
>
> 0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,1,0,0,0,0,-1,0,0,0,0,0,0,0,0,0,0,0,
>
>
> 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,-1,-1,-1,-1,0,0,0,0,1,
>
>
> 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
>
>
> 0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
>
>
> 0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,0,0,0,0,
>
>
> 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,0,0,0,
>
>
> 0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,0,0,
>
>
> 0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,0,
>
>
> 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
>
>
> 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
>
>
> 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
>
>
> 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
>
>
> 0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,0,0,0,0,0,0,1
>
> ),ncol=33,byrow=TRUE)
>
> f.dir <-
c("=","=","=","=","=","=","=","=","=","=","=",
>
"=","=","=","=","=","=","=","=","=","=","=","=",
>
"=","=","=","=","=","=","=","=","=","=")
>
> f.rhs <- c(-52,0,54,0,2,0,0,0,0,0,0,0,0,0)
>
> lp.out <- matrix( ncol = 2, nrow = 13 )
>
> for(ii in 1:amount){ # - simple iterate the same
>
> for( i in 1:13 ){
>
> f.obj[ w[i] ] <- 1
>
> lp.out[ i, 1 ] <- lp("min", f.obj, m, f.dir, f.rhs)$objval
>
> lp.out[ i, 2 ] <- lp("max", f.obj, m, f.dir, f.rhs)$objval
>
> f.obj <- rep( 0, 33 )
>
> }
>
> }
> lp.out
>
> }
>
> ______________________________________________
> 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
--
Arne Henningsen
Department of Agricultural Economics
University of Kiel
Olshausenstr. 40
D-24098 Kiel (Germany)
Tel: +49-431-880 4445
Fax: +49-431-880 1397
ahenningsen at agric-econ.uni-kiel.de
http://www.uni-kiel.de/agrarpol/ahenningsen/