CUDA-动态共享内存触发推力::: system :: system_error

Question

I just started to learn CUDA programming via Udacity. I got the following error even when trying to use dynamic shared memeory.

CUDA error at: main.cpp:55
invalid argument cudaGetLastError()
terminate called after throwing an instance of thrust::system::system_error'
what():  unload of CUDA runtime failed

We are unable to execute your code. Did you set the grid and/or block size correctly?

I searched quite a lot but still had no clue where goes wrong here. Interestingly if I change the last two lines to

    compact_kernel<<<numBlocks, numThreadsPerBlock, sizeof(int)*1000>>>(d_inputVals, d_inputPos, d_outputVals, d_outputPos, numElems, 0);   
    compact_kernel<<<numBlocks, numThreadsPerBlock, sizeof(int)*1000>>>(d_inputVals, d_inputPos, &d_outputVals[numElems/2], &d_outputPos[numElems/2], numElems, 1); 

, no error was thrown when running the code. However, it does not make sense since the space for dynamic memory allocation should not be limited to constant. Maybe it is not my code but the settings on Udacity? The code I wrote is below. Any help would be greatly appreciated.

__global__ void compact_kernel(unsigned int* const d_inputVals,
    unsigned int* const d_inputPos,
    unsigned int* const d_outputVals,
    unsigned int* const d_outputPos,
    const size_t numElems,
    const size_t refBit)
{
    const size_t tid = blockIdx.x * blockDim.x + threadIdx.x;

    // predicate
    const bool predicate = (d_inputVals[tid] & 1) == refBit;
    extern __shared__ int s[];   
}

void your_sort(unsigned int* const d_inputVals,
    unsigned int* const d_inputPos,
    unsigned int* const d_outputVals,
    unsigned int* const d_outputPos,
    const size_t numElems)
{ 
    const size_t numBlocks = numElems/512;
    const size_t numThreadsPerBlock = 256;
    compact_kernel<<<numBlocks, numThreadsPerBlock, sizeof(int)*numElems>>>(d_inputVals, d_inputPos, d_outputVals, d_outputPos, numElems, 0);   
    compact_kernel<<<numBlocks, numThreadsPerBlock, sizeof(int)*numElems>>>(d_inputVals, d_inputPos, &d_outputVals[numElems/2], &d_outputPos[numElems/2], numElems, 1); 

}`

EDIT: The value of numElems is 220480. Is this number too big for dynamic memory allocation?

Answer 1

根据编程指南，对于所有当前CUDA设备， 每个线程块的共享内存限制为48 KB 。