CUDA 8使用-std = gnu ++ 11编译错误 [英] CUDA 8 compile errors with -std=gnu++11

问题描述

我正在尝试将一些代码转换为使用CUDA，并且我认为我遇到兼容性问题。我们使用CMake。这些是我使用的gcc和CUDA版本：

I am trying to convert some code to use CUDA, and I think I am having compatibility issues. We use CMake. These are the gcc and CUDA versions that I use:

$ gcc --version
gcc (Ubuntu 5.4.0-6ubuntu1~16.04.5) 5.4.0 20160609
Copyright (C) 2015 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

$ nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2016 NVIDIA Corporation
Built on Tue_Jan_10_13:22:03_CST_2017
Cuda compilation tools, release 8.0, V8.0.61

在CMakeLists.txt文件中，设置为：

In the CMakeLists.txt file, this is set:

SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=gnu++11")

编译时，出现此错误：

/usr/lib/gcc/x86_64-linux-gnu/5/include/stddef.h(436): error: identifier "nullptr" is undefined

/usr/lib/gcc/x86_64-linux-gnu/5/include/stddef.h(436): error: expected a ";"

/usr/include/x86_64-linux-gnu/c++/5/bits/c++config.h(200): error: expected a ";"

/usr/include/c++/5/exception(63): error: expected a ";"

/usr/include/c++/5/exception(68): error: expected a ";"

/usr/include/c++/5/exception(76): error: expected a ";"

/usr/include/c++/5/exception(83): error: expected a ";"

/usr/include/c++/5/exception(93): error: expected a "{"
...

我发现此讨论，表示将C ++标志更改为 -std = c ++ 98 ，但这不适用于该项目，因为它广泛使用了C ++ 11码。有解决方案吗？它似乎不是一个唯一的问题，但是我似乎找不到解决方法...

I found this discussion, which says to change the C++ flag to -std=c++98 , however this won't do for the project as it makes extensive use of C++11 code. Is there a solution for this? It doesn't appear to be an unique problem, but I can't seem to find a solution...

这是我运行<$ c $时的输出c> cmake ，所以我知道它确实在使用gcc 5.4，而不是其他版本。

This is the output when I run cmake, so I know it is indeed using gcc 5.4 and not another version.

-- The C compiler identification is GNU 5.4.0
-- The CXX compiler identification is GNU 5.4.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Looking for pthread_create
-- Looking for pthread_create - not found
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE  
-- Found CUDA: /usr/local/cuda (found version "8.0") 
-- Found OpenCV: /usr/local (found version "3.2.0") 
-- Found PythonLibs: /usr/lib/x86_64-linux-gnu/libpython2.7.so (found suitable version "2.7.12", minimum required is "2.7") 
-- Found PythonInterp: /usr/bin/python2.7 (found version "2.7.12") 
-- Boost version: 1.58.0
-- Found the following Boost libraries:
--   python
-- Configuring done
-- Generating

推荐答案

如果您想要更新版本的CMake，则更便携的解决方案是

If you are willing to require a more recent version of CMake, the more portable solution is

cmake_minimum_required(VERSION 3.1.3)
set(CMAKE_CXX_STANDARD 11)
set(CMAKE_CXX_STANDARD_REQUIRED ON)  # not necessary, but encouraged

我大约90％确信它与< a href = https://cmake.org/cmake/help/v3.1/variable/CMAKE_CXX_STANDARD.html rel = nofollow noreferrer> v3.1.3 。确保在创建任何特定目标之前先设置（例如， cuda_add_library 或 cuda_add_executable ）。

I'm about 90% sure it came in with v3.1.3. Make sure you set this before you create any specific targets (e.g., cuda_add_library or cuda_add_executable).

还请注意，如果您使用的是传统的 find_package（CUDA）和 cuda_add_library 等，您可能还希望

Note also that if you are using traditional find_package(CUDA) and cuda_add_library etc, you will probably also want

set(CUDA_PROPAGATE_HOST_FLAGS ON)

否则， nvcc 默认情况下不会继承这些定义。

Otherwise nvcc does not by default inherit these definitions.

这篇关于CUDA 8使用-std = gnu ++ 11编译错误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！