Is there a NVCC auto include ?

Accelerated Computing CUDA CUDA Programming and Performance
1

Hi,

I’m currently facing a strange problem, i have to use the function

cudaBindTexture(0,plane_R,cudaPic->plan0_cuda,&cuda_picture_float_array.planTexDesc,width * height * sizeof(float));

with visual 2010 intellisense show no problem, but it doesn’t compile i’m getting

error : no instance of overloaded function “cudaBindTexture” matches the argument list

if i remove

&cuda_picture_float_array.planTexDesc

intellisense underline and say there’s an error but it actually compiles fine.

what’s really strange is if i remove

#include “cuda_runtime_api.h”

everything compile still fine, it’s like nvcc auto include what he need automaticaly except for cudaBindTexture he pick the wrong function and say no overloaded function available.

I use cuda 4.1 , and i ask you if someone ran across the same problem or know how to resolve mine ?

2

ok i come back to say i saw the mistake

here it is :

cudaBindTexture(0,plane_R,p0in,1920*1080 * sizeof(float));

cudaChannelFormatDesc planTexDesc = cudaCreateChannelDescHalf();

cudaBindTexture (0, &plane_R, p0in, &planTexDesc, 1920*1080 * sizeof(float));

One version take a reference for the texture and the other doesn’t…