CODEWITHSUNDEEP
c++rprofilingrcpp
Tim
Tim

Reputation: 99468

Specifying Google perftools to g++ for using C++ in R

I installed Google perftools (google-perftools 1.7-1ubuntu1), and add -lprofiler to PKG_LIBS in R, when compiling the C++ code.

library(RcppArmadillo)
library(Rcpp)

Sys.setenv("PKG_CXXFLAGS"="-fopenmp")
Sys.setenv("PKG_LIBS"="-fopenmp -lprofiler")

sourceCpp('my.cpp')

The output is:

/usr/bin/ld: cannot find -lprofiler
collect2: ld returned 1 exit status
make: *** [sourceCpp_17496.so] Error 1
g++ -I/usr/share/R/include -DNDEBUG    -I"/usr/local/lib/R/site-library/Rcpp/include" -I"/usr/local/lib/R/site-library/RcppArmadillo/include"   -fopenmp -fpic  -g -O2 -fstack-protector --param=ssp-buffer-size=4 -Wformat -Wformat-security -Werror=format-security -D_FORTIFY_SOURCE=2 -g  -c my.cpp -o my.o g++ -shared -Wl,-Bsymbolic-functions -Wl,-z,relro -o sourceCpp_17496.so my.o -llapack -lblas -lgfortran -lm -lquadmath -fopenmp -fopenmp -lprofiler -L/usr/lib/R/lib -lR 
Error in sourceCpp("my.cpp") : 
  Error 1 occurred building shared library.

WARNING: The tools required to build C++ code for R were not found.

Please install GNU development tools including a C++ compiler.

Even if I run g++ in bash, I get the same error

$ g++ -shared -Wl,-Bsymbolic-functions -Wl,-z,relro -o sourceCpp_17496.so my.o -llapack -lblas -lgfortran -lm -lquadmath -fopenmp -fopenmp -lprofiler -L/usr/lib/R/lib -lR 
/usr/bin/ld: cannot find -lprofiler
collect2: ld returned 1 exit status

I wonder why -lprofiler doesn't specify google perfotools? How can I solve the problem? Thanks!

My g++ is g++ (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3.

Upvotes: 1

Views: 1406

Answers (1)

Dirk is no longer here
Dirk is no longer here

Reputation: 368389

Quick ones:

  1. Do you actually have the package libgoogle-perftools-dev installed? Ie do you have shared and static libraries /usr/lib/libprofiler.* ? This is the mother of all development FAQs: you need package libfoo to run code against foo, and package libfoo-dev to compile against foo.

  2. I have old working examples in the slides from my 'HPC with R' talks from a few years ago; those should all work as is.

Upvotes: 8

Related Questions