stl debug support in cuda

Hallo, I would like to debug my CUDA program using gcc macros -D_GLIBCXX_DEBUG -D_GLIBCXX_DEBUG_PEDANTIC. In compilation phase by nvcc I am using syntax: nvcc -Xcompiler -D_GLIBCXX_DEBUG -Xcompiler -D_GLIBCXX_DEBUG_PEDANTIC. But unfortunately, the compiler returns errors at points which are correctly compiled without these macros. I include the error massages of the compiler:

nvcc -Xcompiler -D_GLIBCXX_DEBUG -Xcompiler -D_GLIBCXX_DEBUG_PEDANTIC -Xcompiler -fopenmp -Xcompiler -rdynamic -lineinfo -dc --ptxas-options=-v --maxrregcount=41 down_kernel_fce.cu -arch sm_20 -use_fast_math -prec-div=false
/usr/include/c++/4.6/x86_64-linux-gnu/./bits/c++config.h(223): error: expected an identifier

/usr/include/c++/4.6/x86_64-linux-gnu/./bits/c++config.h(223): error: inline specifier allowed on function declarations only

/usr/include/c++/4.6/x86_64-linux-gnu/./bits/c++config.h(223): error: expected a ";"

down_vars_kernel.hpp(108): error: vector is not a template

down_vars_kernel.hpp(129): error: vector is not a template

down_vars_kernel.hpp(129): error: identifier "vector" is undefined

down_vars_kernel.hpp(129): error: type name is not allowed

down_vars_kernel.hpp(129): error: expected an identifier

down_vars_kernel.hpp(131): error: type name is not allowed

down_vars_kernel.hpp(131): error: expected an identifier

10 errors detected in the compilation of "/tmp/tmpxft_00003a44_00000000-6_down_kernel_fce.cpp1.ii".

What should I do to make these macros working with CUDA?

Thanks in advance,

Dalibor