Duplicate kernel parameters not legal?

Accelerated Computing CUDA CUDA Programming and Performance
1

Why do these two kernels/calls not provide the same answer?

func call

ComplexMul (d_input, d_filter, n_elems);
.
.
.
// Complex multiplication
static __global__ void ComplexMul(cuComplex *a, cuComplex *b, int size)
{
    const int numThreads = blockDim.x * gridDim.x;
    const int threadID = blockIdx.x * blockDim.x + threadIdx.x;

    for (int i = threadID; i < size; i += numThreads)
    {
        a[i].x = a[i].x*b[i].x-a[i].y*b[i].y;
        a[i].y = a[i].x*b[i].y+a[i].y*b[i].x;
    }
}

ComplexMul (d_input, d_filter, d_input, n_elems);
.
.
.


// Complex multiplication
static __global__ void ComplexMul(cuComplex *a, cuComplex *b, cuComplex *c, int size)
{
    const int numThreads = blockDim.x * gridDim.x;
    const int threadID = blockIdx.x * blockDim.x + threadIdx.x;

    for (int i = threadID; i < size; i += numThreads)
    {
        c[i].x = a[i].x*b[i].x-a[i].y*b[i].y;
        c[i].y = a[i].x*b[i].y+a[i].y*b[i].x;
    }
}
2

The first kernel is clobbering a[i].x and then trying to access it in the next statement. You’ll get the updated value, not the original value.

3

User error, as usual. thanks for pointing it out.