similar to: R-alpha: S_alloc

(Sorry for the cross-post--- I wasn't sure which list is more appropriate...) Hi everyone, I've run into segfaults when using my randomForest package on large dataset (e.g., 100 x 15200) and large number of trees (e.g., ntree=7000 and mtry=3000). I'm wondering if anyone can give me some hints on where to look for the problem. The randomForest package mainly consists of two things:

(Sorry for the cross-post--- I wasn't sure which list is more appropriate...) Hi everyone, I've run into segfaults when using my randomForest package on large dataset (e.g., 100 x 15200) and large number of trees (e.g., ntree=7000 and mtry=3000). I'm wondering if anyone can give me some hints on where to look for the problem. The randomForest package mainly consists of two things:

The current version of S_alloc in src/main/memory.c is char *S_alloc(long nelem, int eltsize) { unsigned int i, size; char *p = R_alloc(nelem, eltsize); for(i=0 ; i<size; i++) p[i] = 0; return p; } which segfaults because `size' is not initialized. I am not what the right fix is, adding size = nelem * eltsize; before the loop seems to work. As an aside ... I think the seed*

There is a problem with 'default argument evaluation' when I use an existing function name as argument name : sintest <- function(x, y = 2, sin= sin(pi/4)) { ## Purpose: Test of "default argument evaluation" ## -------- Fails for R-0.49. Martin Maechler, Date: 9 May 97. c(x=x, y=y, sin=sin) } ## R-0.49: R> sintest(1) ##> Error in sintest(1) : recursive

The .Options vector had been introduced a while ago after my suggestion (see Ross's E-mail below). .Options$digits is used be default in several print methods (eg print.lm), however, deparse(.) e.g., uses options()$width, and not .Options$width. Another problem is that .Options is still not in the documentation (on-line help). Before one could add it there, we'd need ``the

Evaluate the following example to get the behavior : --- anybody: patch ? --- tst.i <- function(x) { if(missing(x)) return(invisible()) else if(!is.numeric(x)) stop("x must be numeric") ## else invisible((x+3)^2) } tst.i()#-- should NOT print anything !! print(mode(tst.i()))#--gives "NULL" both in S-plus 3.4 and R 0.60 tst.i(1)# nothing (ok in R and

I see some code in R to attach finalizers to external pointer references (Register[C]Finalizer). Anyone have an example of how to code the finalizer? R_RegisterCFinalizer accepts a C function, but I can't see how to get it to operate on the pointer, since the pointer is not passed to the function when its called. RegisterFinalizer takes an R closure, but how is it called from R (tried

I am trying to write a C function to create a vector of integers that can be used by the R calling function. I do not know the size of the vector in the R calling function. (Well, actually, I have an upper limit on the size, but that is so large that R cannot allocate it. What I'm doing in the function is to do a sieving procedure, and the result will be small enough to fit into my

Prof. Tierney, Thanks very much for the info. Why does the loop work if I move the assignment out of the condition? E.g., the following works: i <- 0 while(i < 5) { i <- i + 1 if(i < 3) next print(i) } Regards, Andy > -----Original Message----- > From: Luke Tierney [mailto:luke at nokomis.stat.umn.edu] > Sent: Tuesday, October 02, 2001 3:36 PM > To: Liaw, Andy

Running on an Alpha machine which reports (uname -a) OSF1 bsdx01.bs.ehu.es V4.0 878 alpha and using the binary distribution put together by Albrecht Gebhardt (in http://cran.at.r-project.org/bin/osf/osf4.0/tar/alpha_ev5/) I obtain core dumps whenever I try to use package rpart. I have R REMOVE'd the rpart package, downloaded the source rpart_1.0-7.tar from CRAN and

I'm curious if any of the core R developers have considered the possibility of hosting R (v2?) on the parrot common runtime environment. Perl6 will generate byte-code for parrot, as will some future version of python. I can imagine both drawbacks and advantages. Some advantages would be fast byte-code execution and freely mixing perl, python and R modules. Anyone looked into this? Tim

Can a C function wrapped with .Call return void? Or must it return a valid SEXP? Thnx. Tim -- Timothy H. Keitt Department of Ecology and Evolution State University of New York at Stony Brook Stony Brook, New York 11794 USA Phone: 631-632-1101, FAX: 631-632-7626 http://life.bio.sunysb.edu/ee/keitt/ -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel mailing

By my reading, R has a hard limit on INT_MAX bytes for vector allocations. On the sparcv9 architecture, INT_MAX is a 32 bit quantity, even though pointers can be 64 bits. Has any thought been given to use of > 2GB of virtual memory on systems like this? -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read

I wonder if it was possible (and desireable) to implement means to protect data objects (and functions) from overwriting them. So for instance: > x <- 5 > protect(x) > x <- 8 Error: x is read only Or would that be against the philosophy of R to be compatible to S?! Just a RFC ;-) Yours, Christoph. ps. Your probably know the situations that let you think of such a feature

luke's demonstration of external references, at www.stat.umn.edu/~luke/file_0.0-0.tar.gz (accessible through the developer page) causes a segfault on solaris 2.8 as soon as library(file) is attempted. it runs fine on RH 7.2. the segfault on solaris occured for R 1.5.1 and 1.6 (8-25 image) a quick poke at gdb (which i can only get running for R 1.6, though this bug report is for 1.5.1)

Hi! I've been able to configure and make R (1,4,0 linux) except for the gnome support. I've checked all packages and libs mentioned in R-admin.pdf (and by Luke Tierny) to make sure that I have them and even using the following: ./configure --with-gnome --with-gnome-includes=/opt/gnome/include --with-gnome-libs=/opt/gnome/lib --with-libglade-config=opt/gnome/lib I still get: R is now

Hi, R-1.4.0 on ES40 under OSF 5.1. I use gcc version 3.0.3 During gnu make I have the message : dumping R code in package `methods' Error in dyn.load(x, as.logical(local), as.logical(now)) : unable to load shared library"/usr/local/R1.4.0/library/methods/libs/methods.so": dlopen: /usr/local/R-1.4.0/library/ctest/libs/ctest.so: symbol "dansari" unresolved Execution

I like Thomas' example. Kurt can you include it in the FAQ. Recently I have been doing a lot of programming that has required either optimization or zero-finding. Unfortunately most of the functions need some arguments to be optimized over and have other parameters that depend on the data but are fixed with respect to optimization. With R scoping rules this is a trivial problem; simply

I like Thomas' example. Kurt can you include it in the FAQ. Recently I have been doing a lot of programming that has required either optimization or zero-finding. Unfortunately most of the functions need some arguments to be optimized over and have other parameters that depend on the data but are fixed with respect to optimization. With R scoping rules this is a trivial problem; simply

Has anyone experimented with the Network Common Data Format (netcdf) described at http://www.unidata.ucar.edu/packages/netcdf for archiving data? Is this format widely used outside of atmospheric research? -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info",