使用用户自定义的目录安装cuda，报环境不匹配的警告

[yifanfeng97]

Environment

Hardware Environment(Ascend/GPU/CPU):

/device gpu

Software Environment:

• MindSpore version (source or binary): 1.1.0
• Python version (e.g., Python 3.7.5): 3.7.6
• OS platform and distribution (e.g., Linux Ubuntu 16.04): Ubuntu 18.04
• GCC/Compiler version (if compiled from source): 7.5.0 用户自己安装的CUDA环境截图：

用户自己的home下的.bashrc中的配置：

系统的CUDA环境：

Describe the current behavior

import mindspore 报出警告说环境不匹配：[WARNING] ME(39922:140551200737088,MainProcess):2021-01-23-15:05:55.109.91 [mindspore/_check_version.py:96] MindSpore version 1.1.0 and cuda version 9.0 does not match, reference to the match info on: https://www.mindspore.cn/install

Describe the expected behavior

希望mindspore能正确检测系统的cuda环境的版本，从而不报出这种警告。

Steps to reproduce the issue

1. 系统管理员在系统层面安装了CUDA 9.0
2. 用户在自己的目录用非root权限安装了CUDA 10.1
3. 用户通过修改.bashrc文件，将自己安装的CUDA 10.1的环境，放到 LD_LIBRARY_PATH 中，CUDA的bin至于PATH的最前面
4. 命令行输入python
5. 输入import mindspore
6. 报出环境不匹配的警告：[WARNING] ME(39922:140551200737088,MainProcess):2021-01-23-15:05:55.109.91 [mindspore/_check_version.py:96] MindSpore version 1.1.0 and cuda version 9.0 does not match, reference to the match info on: https://www.mindspore.cn/install

Related log / screenshot

Special notes for this issue

我简单查阅了一下报错的源码的位置：https://gitee.com/mindspore/mindspore/blob/master/mindspore/_check_version.py#L86 发现该函数依赖写死的cuda路径导致了该警告，也就是在https://gitee.com/mindspore/mindspore/blob/master/mindspore/_check_version.py#L48 希望能借鉴pytorch或者tensorflow的check version函数，不依赖写死的cuda安装路径去检查版本。 实际上我的环境下面pytorch和tensorflow都能正常check cuda的版本。

[cxl-ll]

Hello.thank you for your feedback. We plan to solve it in the next version. We will let you know when we finish.

[ytgui]

That because the mindspore boys know nothing about ubuntu x86 and x86_64. Temporary solution: use ln -s command to create a softlink, this could cheat cmake and build mindspore successful without changing original CMakeList.txt