Is there any way I can call CUDA runtime function calls such as
cudaMemcpy(...);
in a .cpp file, compiled with a regular C++ compiler?
NVCC is a compiler driver which works by invoking all the necessary tools and compilers like cudacc, g++, cl, etc. NVCC can output either C code (CPU Code) that must then be compiled with the rest of the application using another tool or PTX or object code directly.
cu files, since NVCC 2. x actually compiles these files as c++ files. c++ compilation enforces a different syntax for function symbols in object code, which resolved in linking errors of the type “undefined reference” when linking .
EDIT: There was an example here but it's not longer found, but most of the example was copied below.
The caller C (but could be C++)
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <cuda.h>
extern void kernel_wrapper(int *a, int *b);
int main(int argc, char *argv[])
{
int a = 2;
int b = 3;
kernel_wrapper(&a, &b);
return 0;
}
The Callee (CUDA)
__global__ void kernel(int *a, int *b)
{
int tx = threadIdx.x;
switch( tx )
{
case 0:
*a = *a + 10;
break;
case 1:
*b = *b + 3;
break;
default:
break;
}
}
void kernel_wrapper(int *a, int *b)
{
int *d_1, *d_2;
dim3 threads( 2, 1 );
dim3 blocks( 1, 1 );
cudaMalloc( (void **)&d_1, sizeof(int) );
cudaMalloc( (void **)&d_2, sizeof(int) );
cudaMemcpy( d_1, a, sizeof(int), cudaMemcpyHostToDevice );
cudaMemcpy( d_2, b, sizeof(int), cudaMemcpyHostToDevice );
kernel<<< blocks, threads >>>( a, b );
cudaMemcpy( a, d_1, sizeof(int), cudaMemcpyDeviceToHost );
cudaMemcpy( b, d_2, sizeof(int), cudaMemcpyDeviceToHost );
cudaFree(d_1);
cudaFree(d_2);
}
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With