whoami
April 17, 2009, 7:36pm
1
I am a beginer about CUDA, just want to follow some examples in the Nvidia Instruction. So I try this sample program:
#include <stdio.h>#include <stdlib.h>#include <cuda.h>
global void vecAdd(float* A, float* B, float* C)
int main()
A[0]=5.0f;
vecAdd<<<1, N>>>(A, B, C);
But it doesn’t work. Both C[0] and C[1] print to be 0. I think this is some basic problem I need to figure out. Thanks for everybody who can help.
I am a beginer about CUDA, just want to follow some examples in the Nvidia Instruction. So I try this sample program:
include
include
include
global void vecAdd(float* A, float* B, float* C)
{
int i = threadIdx.x;
C[i]=A[i]+B[i];
}
int main()
{
float A[2];
float B[2];
float C[2];
int N=2;
A[0]=5.0f;
A[1]=1.0;
B[0]=10.0f;
B[1]=3.0;
vecAdd<<<1, N>>>(A, B, C);
printf(“A=%f, B=%f,C= %f\n”, A[0], B[0], C[0]);
printf(“A2=%f, B2=%f, C2=%f\n”, A[1], B[1], C[1]);
}
But it doesn’t work. Both C[0] and C[1] print to be 0. I think this is some basic problem I need to figure out. Thanks for everybody who can help.
The kernel “thinks” that A, B and C point to device memory and therefore interpretes these adresses as adresses of device mem. But actually A, B, C are pointers to arrays in host mem.