how to compile and run the Dynamic Parallelism code (DCP ) sample code in visuall studio 2005??

Accelerated Computing CUDA CUDA Setup and Installation
1

I have a card GTX 780 which is compute capability 3.5 . It runs the basic program properly but whenever I tried to build the sample code of DCP for sample provided with cuda . It generates error. as soon as I add cuddevrt.lib and change the relocateable devide code to yes It generates error even in simple program as

Error	2	error LNK2019: unresolved external symbol ___cudaRegisterLinkedBinary_38_tmpxft_00000dd8_00000000_8_new_cpp1_ii_g_uids referenced in function "void __cdecl __sti____cudaRegisterAll_38_tmpxft_00000dd8_00000000_8_new_cpp1_ii_g_uids(void)" (?__sti____cudaRegisterAll_38_tmpxft_00000dd8_00000000_8_new_cpp1_ii_g_uids@@YAXXZ)	new.cu.obj

and

Error	3	fatal error LNK1120: 1 unresolved externals	C:\Users\Admin\Documents\Visual Studio 2005\Projects\dynamic_parallilism\Debug\dynamic_parallilism.exe

the visual studio command is

1>------ Rebuild All started: Project: dynamic_parallilism, Configuration: Debug Win32 ------
1>Deleting intermediate and output files for project 'dynamic_parallilism', configuration 'Debug|Win32'
1>Compiling with CUDA Build Rule...
1>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\bin\nvcc.exe"     -gencode=arch=compute_35,code=\"sm_35,compute_35\" -gencode=arch=compute_35,code=\"sm_35,compute_35\"  --machine 32 -ccbin "C:\Program Files (x86)\Microsoft Visual Studio 8\VC\bin"    -Xcompiler "/EHsc /W3 /nologo /O2 /Zi   /MT  "  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v5.0\include" -maxrregcount=0  -rdc=true --compile -o "Debug/new.cu.obj" new.cu  
1>nvcc : warning : nvcc support for MSVC 8.0 and earlier has been deprecated and is no longer being maintained
1>new.cu
1>tmpxft_00000dd8_00000000-5_new.cudafe1.gpu
1>tmpxft_00000dd8_00000000-10_new.cudafe2.gpu
1>new.cu
1>tmpxft_00000dd8_00000000-5_new.cudafe1.cpp
1>tmpxft_00000dd8_00000000-15_new.ii
1>Compiling manifest to resources...
1>Linking...
1>new.cu.obj : error LNK2019: unresolved external symbol ___cudaRegisterLinkedBinary_38_tmpxft_00000dd8_00000000_8_new_cpp1_ii_g_uids referenced in function "void __cdecl __sti____cudaRegisterAll_38_tmpxft_00000dd8_00000000_8_new_cpp1_ii_g_uids(void)" (?__sti____cudaRegisterAll_38_tmpxft_00000dd8_00000000_8_new_cpp1_ii_g_uids@@YAXXZ)
1>C:\Users\Admin\Documents\Visual Studio 2005\Projects\dynamic_parallilism\Debug\dynamic_parallilism.exe : fatal error LNK1120: 1 unresolved externals
1>Build log was saved at "file://c:\Users\Admin\Documents\Visual Studio 2005\Projects\dynamic_parallilism\dynamic_parallilism\Debug\BuildLog.htm"
1>dynamic_parallilism - 2 error(s), 1 warning(s)
========== Rebuild All: 0 succeeded, 1 failed, 0 skipped ==========
2

The first CUDA toolkit that supports dynamic parallelism is CUDA 5.0. VS 2005 is not a supported environment for CUDA 5.0. Note the warning in your output:

1>nvcc : warning : nvcc support for MSVC 8.0 and earlier has been deprecated and is no longer being maintained

You can also verify this through the CUDA 5.0 release notes, which only list VS 2008 and VS 2010 as supported environments.

3

and when Tested in visual studio 2010, as soon as I add cudadevrt.lib and changed the Generate Relocateable Device code to yes, it compiles but the kernel is not executed.

4

Then you should add proper cuda error checking to your code, and get the actual error that is returned after the kernel call.