Error when use SyncThreadsCount function

Sep 27, 2014 at 8:24 PM
Dear all,,,

I got this exception "SyncThreadsCount is not supported for compute capability 1.3."
int y = thread.blockIdx.y * thread.blockDim.y + thread.threadIdx.y;
            int x = thread.blockIdx.x * thread.blockDim.x + thread.threadIdx.x;
            int i = ((thread.gridDim.x * thread.blockDim.x) * y) + x;
            int offset = 0;
            bool numSwap;

           
            do
            {
                i += offset;
                numSwap = false;
                int d0=data[i];
                int d1=data[i + 1];

                int v0 = indexs[i];
                int v1 = indexs[i + 1];
               
                if (i > max)
                    return;
                
                if (d0 > d1)
                {
                    data[i] = d1;
                    data[i+1] = d0;

                    indexs[i] = v1;
                    indexs[i + 1] = v0;
                    numSwap=true;
                }

                if (offset == 0)
                {
                    offset = 1;
                }
                else
                {
                    offset = 0;
                }

            }
            while (thread.SyncThreadsCount(numSwap) != 0) ;
 
please help me in this exception.

I used GPU GT 740M

Regards
Oct 3, 2014 at 10:27 PM
you need to call CudafyTranslator.Cudafy using an overload that takes eArchitecture and that architecture must be greater than 2