cuda error checking 学习笔记

由于发现cuda c++ 的 debug方式和c++ 差别很大,因此打算再开一篇,专门记录一些和error checking 以及debug有关的内容.

Error checks in CUDA code can help catch CUDA errors at their source. There are 2 sources of errors in CUDA source code:

  1. Errors from CUDA **API** calls. For example, a call to `cudaMalloc()` might fail.
  2. Errors from CUDA **kernel** calls. For example, there might be invalid memory access inside a kernel.

All CUDA API calls return a cudaError value, so these calls are easy to check:

1
2    
3    if ( cudaSuccess != cudaMalloc( &fooPtr, fooSize ) )
4        printf( "Error!\n" );
5    

CUDA kernel invocations do not return any value. Error from a CUDA kernel call can be checked after its execution by calling cudaGetLastError():

1fooKernel<<< x, y >>>(); // Kernel call
2if ( cudaSuccess != cudaGetLastError() )
3    printf( "Error!\n" );

以及下面是一个check error的宏…

1#define CHECK_ERROR( err ) \
2  if( err != cudaSuccess ) { \
3    std::cerr << "ERROR: " << cudaGetErrorString( err ) << std::endl; \
4    exit( -1 ); \
5  }

1#define CHECK_LAST_ERROR \
2  { cudaError_t err = cudaGetLastError(); \
3    if( err != cudaSuccess ) { \
4      std::cerr << cudaGetErrorString( err ) << std::endl; \
5      exit( -1 ); \
6    } \
7  }