CUDA kernel stuck

Accelerated Computing CUDA CUDA Programming and Performance
1

Dear All,

     I am having a problem running a cuda kernel. It stucks for long time and I don't have a cluse as to what is causing it. The kernel is given below.

__global__ void Inside_Error_Calculations(

		float *d_proj_cyls_data, 

		unsigned char *d_FGMaps, float *errors, int *sample_points){

float error = 0;

  int samples=0;

  int FGMap_stride = (640*480/8)+1;  

  int error_sample_index=blockIdx.x*blockDim.x + threadIdx.x;

  int my_index;

  int proj_data_stride = InError_Threads*InError_Blocks*8; // 64 blocks, 64 threads per block

  int proj_data_size   = 4096*10*8;

  int cnt = proj_data_size/proj_data_stride; 

for (int map = 0; map < 4; map++){

int FGMap_index = map*FGMap_stride;

    my_index = blockIdx.x*8*blockDim.x + threadIdx.x*8 + map*proj_data_size;

	

    for (int i = 0; i < cnt; i++){

       Inside_Error(&d_proj_cyls_data[my_index], &d_FGMaps[FGMap_index], &error, &samples);

       my_index += proj_data_stride;

       errors[error_sample_index]        = error;

       sample_points[error_sample_index] = samples;	

       error=0; samples=0;

       error_sample_index += InError_Threads*InError_Blocks; // 64 blocks, 64 threads per block

    }     

  }	

}

When I set cnt = 1 and run 64x64 threads, it runs ok, but when cnt is not one it just stucks even when it is 2. I am running the kernel on Fermi (GTX 480) with cuda run time 3.2. Can some one tell me as to what is going wrong.

Best regards