2

I have started using JCuda.When I compile the program I get "jcuda.cuda exception cuda_error_launch_failed" error.I have followed the steps given in the installation guide but doesn't work.Very few resources are available online.So,Please help me.This is my code:

extern "C"
__global__ void reduce(float *g_idata, float *g_odata, unsigned int n)
{
    extern __shared__ float sdata[];
    unsigned int tid = threadIdx.x;
    unsigned int i = blockIdx.x*blockDim.x*2 + threadIdx.x;
    unsigned int gridSize = blockDim.x*2*gridDim.x;

    float mySum = 0;


    while (i < n)
    {
        mySum += g_idata[i];
        if (i + blockDim.x < n)
            mySum += g_idata[i+blockDim.x];
        i += gridSize;
    }

    sdata[tid] = mySum;
    __syncthreads();

    if (blockDim.x >= 512) { if (tid < 256) { sdata[tid] = mySum = mySum + sdata[tid + 256]; } __syncthreads(); }
    if (blockDim.x >= 256) { if (tid < 128) { sdata[tid] = mySum = mySum + sdata[tid + 128]; } __syncthreads(); }
    if (blockDim.x >= 128) { if (tid <  64) { sdata[tid] = mySum = mySum + sdata[tid +  64]; } __syncthreads(); }

    if (tid < 32)
    {
        volatile float* smem = sdata;
        if (blockDim.x >=  64) { smem[tid] = mySum = mySum + smem[tid + 32]; }
        if (blockDim.x >=  32) { smem[tid] = mySum = mySum + smem[tid + 16]; }
        if (blockDim.x >=  16) { smem[tid] = mySum = mySum + smem[tid +  8]; }
        if (blockDim.x >=   8) { smem[tid] = mySum = mySum + smem[tid +  4]; }
        if (blockDim.x >=   4) { smem[tid] = mySum = mySum + smem[tid +  2]; }
        if (blockDim.x >=   2) { smem[tid] = mySum = mySum + smem[tid +  1]; }
    }

    if (tid == 0)
        g_odata[blockIdx.x] = sdata[0];
}

I have nvidia 460 gtx gpu and cuda drivers and toolkit installed.

user1633170
  • 11
  • 2
  • 13

1 Answers1

1

Use cudaFree(deviceId); or cudaFree(deviceId); to free memory and try again.Hope this helps you.

user1613360
  • 1,280
  • 3
  • 16
  • 42