A problem about cudaMemcpyAsync and how can I remote control my computer to execute CUDA program

Accelerated Computing CUDA CUDA Programming and Performance
1

A problem about cudaMemcpyAsync and how can I remote control my computer to execute CUDA program

(1)My program sement is as following, but when I execute it,it will show me “invalid argument” problem.I think “Data_in+inumTxdeal_length2-numTxmaxD2first_or_not” when i is not equal to 0 will cause this problem. Who can help me ?

float *Data_in;

b=cudaMallocHost((void**)&Data_in,numTx*(datalength+maxD+order)*2);

a=cudaMemcpyAsync(Data_in_GPU, Data_in+inumTxdeal_length2-numTxmaxD2first_or_not , numTx*(last_length+first_or_not*maxD)2sizeof(float), cudaMemcpyHostToDevice ,stream1[i]);

Fractional_filter<<<dimGrid_Frac, dimBlock_Frac,0,stream1[i]>>>(frac_outI_GPU, frac_outQ_GPU, FB_GPU, FF_GPU, last_length,order,numTx,Data_in_GPU,fradelay_GPU,TapNum,intd
elay_GPU,first_or_not,maxD);

(2)How can I set my computer,so I can remote control my computer to execute CUDA program.

  Thanks Everyone.