When I compile my CUDA code,it said:1 error detected in the compilation of "/tmp/tmpxft_00005ff8_00000000-4_lbm.cpp1.ii" -

When my CUDA code compiled on the basis of command:

NVCC lbm.cu I / usr / local / NVIDIA_GPU_Computing_SDK / CUDALibraries / common / Ink -lm

the following compiler error information:

  file contains lbm.cu:15 in: lbm_kernel.h: 52: 8: Warning: error: #endif instructions token identifier at the end of /usr/local/NVIDIA_GPU_Computing_SDK/CUDALibraries/common/inc/cutil_inline_runtime.h(329) "CURAND_STATUS_DOUBLE_PRECISION_REQUIRED" is undefined detected 1 error In the compilation of "/tmp/tmpxft_00005ff8_00000000-4_lbm.cpp1.ii"   
 This is the entire content I have never seen it, what can I do during the other codes of my code?   
  
 A  Error  and  Warning . Error is almost certainly due to the listed status:  
  error: identifier "CURAND_STATUS_DOUBLE_PRECISION_REQUIRED" undefined   
 If you want to know you Why this error is occurring, it will be necessary to inspect your code, especially that particular identifier ( CURAND_STATUS_DOUBLE_PRECISION_REQUIRED ) may be due to that line, possibly due to this fact That you 't  #include "curand.h"  your In the program, which provides a definition of a particular identifier (enum value) in  
 The warning is due to an unusual creation.  
  __i devices__ int ind (int x, int y, int z, int k, int4 dim); __device__ int ind (int x, int y, int dim_x); #endif _LBM_KERNEL_H   
 That last line might be:  
  #endif // _LBM_KERNEL_H