"unsigned int*"类型的参数与"size_t*"类型的参数不兼容

Question

我用c ++在cuda中有这个代码:

// Variables
float        *query_dev;
float        *ref_dev;
float        *dist_dev;
int          *ind_dev;
cudaArray    *ref_array;
cudaError_t  result;
size_t       query_pitch;
size_t       query_pitch_in_bytes;
size_t       ref_pitch;
size_t       ref_pitch_in_bytes;
size_t       ind_pitch;
size_t       ind_pitch_in_bytes;
size_t       max_nb_query_traited;
size_t       actual_nb_query_width;
unsigned int memory_total;
unsigned int memory_free;

// Check if we can use texture memory for reference points
unsigned int use_texture = ( ref_width*size_of_float<=MAX_TEXTURE_WIDTH_IN_BYTES && height*size_of_float<=MAX_TEXTURE_HEIGHT_IN_BYTES );

// CUDA Initialisation
cuInit(0);

// Check free memory using driver API ; only (MAX_PART_OF_FREE_MEMORY_USED*100)% of   memory will be used

CUcontext cuContext;
CUdevice  cuDevice=0;
cuCtxCreate(&cuContext, 0, cuDevice);
cuMemGetInfo(&memory_free, &memory_total);

编译时遇到错误:cuMemGetInfo(&memory_free,&memory_total);

错误是:

app.cu(311): error: argument of type "unsigned int *" is incompatible with parameter of type "size_t *"

app.cu(311): error: argument of type "unsigned int *" is incompatible with parameter of type "size_t

311是以下行: cuMemGetInfo(&memory_free, &memory_total);

我不知道这个错误是什么,你对此有什么想法吗？

Answer 1

更改以下行:

unsigned int memory_total;
unsigned int memory_free;

至:

size_t memory_total;
size_t memory_free;

您可能正在尝试在CUDA 3.0之前最初构建的旧代码.

资源

Answer 2

错误说明size_t并且unsigned int是不同的类型,因此您不能将指针传递给期望另一个的函数.

要么改变的类型memory_free,并memory_total以size_t或使用临时size_t变量,然后将值复制到memory_free和memory_total

PS你发布了太多的源代码,请尽量减少你的例子.