第九章中初次介绍原子操作用于归约时编译报错

[zzyuanyi]

输入的编译命令为:nvcc _O3 _DUSE_DP xxx.cu。报错信息如下：9.cu(37): error: no instance of overloaded function "atomicAdd" matches the argument list
argument types are: (real *, real)
atomicAdd(&d_y[0],x[0]);
^
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include\sm_20_atomic_functions.hpp(82): note #3326-D: function "atomicAdd(float *, float)" does not match because argument #1 does not match parameter
static __inline __declspec(device) float atomicAdd(float *address, float val)
^
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include\device_atomic_functions.hpp(224): note #3326-D: function "atomicAdd(unsigned long long *, unsigned long long)" does not match because argument #1 does not match parameter
static __inline __declspec(device) unsigned long long int atomicAdd(unsigned long long int *address, unsigned long long int val)
^
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include\device_atomic_functions.hpp(110): note #3326-D: function "atomicAdd(unsigned int *, unsigned int)" does not match because argument #1 does not match parameter
static __inline __declspec(device) unsigned int atomicAdd(unsigned int *address, unsigned int val)
^
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include\device_atomic_functions.hpp(105): note #3326-D: function "atomicAdd(int *, int)" does not match because argument #1 does not match parameter
static __inline __declspec(device) int atomicAdd(int *address, int val)
GPU型号为3060-laptop，os为win11

[brucefan1983]

书中应该说了，使用双精度的原子加函数时需要用6.0或以上的架构设置。所以编译命令应该需要：

nvcc  -O3 -DUSE_DP -arch=sm_60 xxx.cu  

[zzyuanyi]

十分感谢，问题已解决，多有叨扰，非常抱歉