__CUDA_ARCH__宏的行为 [英] The behavior of __CUDA_ARCH__ macro

问题描述

在宿主代码中，看来 __ CUDA_ARCH __ 宏不会生成不同的代码路径，而是会为当前设备的确切代码路径生成代码.

In the host code, it seems that the __CUDA_ARCH__ macro wont generate different code path, instead, it will generate code for exact the code path for the current device.

但是，如果 __ CUDA_ARCH __ 在设备代码中，它将为编译选项(/arch)中指定的不同设备生成不同的代码路径.

However, if __CUDA_ARCH__ were within device code, it will generate different code path for different devices specified in compiliation options (/arch).

任何人都可以确认这是正确的吗?

Can anyone confirm this is correct?

推荐答案

__ CUDA_ARCH __ 在设备代码中使用时，将带有为其定义的数字，以反映当前正在编译的代码体系结构.

__CUDA_ARCH__ when used in device code will carry a number defined to it that reflects the code architecture currently being compiled.

它不打算在主机代码中使用.从nvcc 手册:

It is not intended to be used in host code. From the nvcc manual:

此宏可用于实现GPU功能，以确定当前为其编译的虚拟体系结构.主机代码(非GPU代码)不得依赖于它.

This macro can be used in the implementation of GPU functions for determining the virtual architecture for which it is currently being compiled. The host code (the non-GPU code) must not depend on it.

因此，主机代码中 __ CUDA_ARCH __ 的用法未定义(至少由CUDA定义).正如@tera在评论中指出的那样，由于宏未在主机代码中定义，因此可以用于区分主机/设备路径，例如，在 __ host__ __device __ 函数定义中.

Usage of __CUDA_ARCH__ in host code is therefore undefined (at least by CUDA). As pointed out by @tera in the comments, since the macro is undefined in host code, it could be used to differentiate host/device paths for example, in a __host__ __device__ function definition.

#ifndef __CUDA_ARCH__
//host code here
#else
//device code here
#endif

这篇关于__CUDA_ARCH__宏的行为的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！