I'm trying to get R running on AIX 5.1 with the native AIX compilers, VisualAge C++ 5.0.2.0 and XL Fortran Compiler 7.1.0.0. R compiled fine, but in running the test cases, the base-Ex.R tests fail in the Choleski decomposition with the following:> x <- matrix(c(1:5, (1:5)^2), 5, 2) > m <- crossprod(x) > Q <- chol(m)Error in chol(m) : NA/NaN/Inf in foreign function call (arg 1) Execution halted I modified the test case to show the intermediate results (and creating m and m1 with the intention of getting beyond this test case) with the following:> x <- matrix(c(1:5, (1:5)^2), 5, 2) > x[,1] [,2] [1,] 1 1 [2,] 2 4 [3,] 3 9 [4,] 4 16 [5,] 5 25> m1 <- crossprod(x) > m1[,1] [,2] [1,] NA 225 [2,] 225 979> m <- t(x) %*% x > m[,1] [,2] [1,] 55 225 [2,] 225 979> Q <- chol(m) > Q[,1] [,2] [1,] 7.416198 30.33899 [2,] 0.000000 7.65150> all.equal(t(Q) %*% Q, m) # TRUE[1] TRUE This gets beyond the failing test case (but does nothing to fix the crossprod) and then the trees test case fails with:> step(fm2)Start: AIC= -152.69 log(Volume) ~ log(Girth) + log(Height) Error in drop1.lm(fit, scope$drop, scale = scale, trace = trace, k = k, : scope is not a subset of term labels Execution halted Any tips on what might be wrong or how to go about debugging these errors? I also tried using GNU/gcc 3.1 and ran into other problems in the test cases. Andy Pierce -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
"Andy Pierce" <apierce at stny.rr.com> writes:> [,1] [,2] > [1,] 1 1 > [2,] 2 4 > [3,] 3 9 > [4,] 4 16 > [5,] 5 25 > > m1 <- crossprod(x) > > m1 > [,1] [,2] > [1,] NA 225 > [2,] 225 979...> This gets beyond the failing test case (but does nothing to fix the crossprod) and > then the trees test case fails with:...> Any tips on what might be wrong or how to go about debugging these > errors?Start with the first one. You're not going to trust a system that miscomputes a crossproduct! You need to figure out exactly how that NA sneaks into m1. That's about as close as I can get since I don't actually know what debuggers you have available. With gdb, I'd set a breakpoint in symcrossprod() inside src/main/array.c. -- O__ ---- Peter Dalgaard Blegdamsvej 3 c/ /'_ --- Dept. of Biostatistics 2200 Cph. N (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk) FAX: (+45) 35327907 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Andy Pierce wrote:> I'm trying to get R running on AIX 5.1 with the native AIX compilers, VisualAge C++ > 5.0.2.0 and XL Fortran Compiler 7.1.0.0. R compiled fine, but in running the test > cases, the base-Ex.R tests fail in the Choleski decomposition with the following:[snip]> Any tips on what might be wrong or how to go about debugging these errors? I also > tried using GNU/gcc 3.1 and ran into other problems in the test cases.Hi, I seem to remember that I had problems running the test cases when using the native fortran libs. I have only tried to compile R with gcc recently. Do you only use the C compiler from gcc or also the Fortran compiler? Can you make sure that the correct libs (gcc's not xlc's) are picked up during linking? -tom -- mailto:vogels at cmu.edu -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._