Hey dudes, here is my code
float *hst_matData2=new int[hst_totalSze];
cudaMemcpy( hst_matData2,dev_matData2, hst_totalSze * sizeof(int),cudaMemcpyDeviceToHost);
calculation on dev_matData2 will be done in kernel but the problem is with cudaMemcpy
how can i do this thing ???
This shouldn’t compile at all:
float *hst_matData2=new int[hst_totalSze];
But I am somewhat confused. You are reporting a cudaMalloc error, but there is no cudaMalloc call in code you posted.
sorry my fault i was in a bit hurry :P
on serious note my apology