Ubuntu安装CUDA8.0+cuDNN7

文章目录

一、准备

1.1 检查是否支持CUDA

输入下面命令查看电脑的NVIDIA型号：

(CCNet36) bit@bit-613:~/下载$ lspci | grep -i nvidia
01:00.0 VGA compatible controller: NVIDIA Corporation GM107 [GeForce GTX 745] (rev a2)
01:00.1 Audio device: NVIDIA Corporation Device 0fbc (rev a1)

显示型号是：​​GeForce GTX 745​​

1.2 查看gcc是否安装

bit@bit-613:~$ gcc --version
gcc (Ubuntu 5.4.0-6ubuntu1~16.04.12) 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.

gcc安装的版本：​​Ubuntu 5.4.0​​

1.3 检查内核版本

(CCNet36) bit@bit-613:~$ uname -r
4.15.0-123-generic

内核版本：​​4.15​​

二、安装显卡驱动

详细过程：
​​​Ubuntu如何查看显卡信息及安装NVDIA显卡驱动​​​​ubuntu 怎么查看显卡型号及安装显卡驱动​​

1、首先，进入ubuntu系统桌面，单击左上bai角的“搜索”选项，du搜索“驱动”，然后打开搜索到的驱动程序，如下图所示，然后进入下一步。

2、接着，完成上述步骤后，将显示计算机每个设备的驱动程序。
如果在其中找到显卡设备，则可以查看显卡并安装驱动程序，点击“应用更改”按钮，如下图所示，然后进入下一步。

3、输入ubuntu系统密码，然后单击“授权”按钮以开始安装驱动程序，如下图所示，然后进入下一步。
4、系统将自动下载并安装驱动程序。
5、安装后单击“重新启动”按钮以重新启动计算机。

6、重新启动计算机后，可以在搜索范围内查看已安装的插件驱动程序。

若不存在显卡，则去官网下载显卡驱动。​​网址地址​​。

7、成功安装驱动后，可以进行查询：

三、安装CUDA

安装之前首先要确认你需要安装的cuda。
​​​TensorFlow+cudnn、cuda、Python的配套关系_包含所有操作系统​​

3.1 CUDA对应的NVIDIA驱动版本对照表

3.2 下载适合的CUDA8.0

下载地址：​​CUDA Toolkit 8.0 - Feb 2017​​​

3.3 安装

1、安装：

sudo dpkg -i cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64 (1).deb
sudo apt-key add /var/cuda-repo-8-0-local-ga2/7fa2af80.pub
sudo apt-get update
sudo apt-get -y install cuda

• 1
• 2
• 3
• 4

2、添加环境变量：

sudo gedit ~/.bashrc

• 1

在文件中追加：

export PATH=/usr/local/cuda-8.0/bin${PATH:+:$PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

• 1
• 2

3.4 验证安装的正确性

1、

cd /usr/local/cuda-8.0/samples/1_Utilities/deviceQuery
 sudo make
 ./deviceQuery

bit@bit-613:/usr/local/cuda-8.0/samples/1_Utilities/deviceQuery$ sudo make
/usr/local/cuda-8.0/bin/nvcc -ccbin g++ -I../../common/inc  -m64    -gencode arch=compute_20,code=sm_20 -gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_52,code=sm_52 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_60,code=compute_60 -o deviceQuery.o -c deviceQuery.cpp
nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
/usr/local/cuda-8.0/bin/nvcc -ccbin g++   -m64      -gencode arch=compute_20,code=sm_20 -gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_52,code=sm_52 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_60,code=compute_60 -o deviceQuery deviceQuery.o 
nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
mkdir -p ../../bin/x86_64/linux/release
cp deviceQuery ../../bin/x86_64/linux/release
bit@bit-613:/usr/local/cuda-8.0/samples/1_Utilities/deviceQuery$ 
bit@bit-613:/usr/local/cuda-8.0/samples/1_Utilities/deviceQuery$ ./deviceQuery 
./deviceQuery Starting...

 CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 1 CUDA Capable device(s)

Device 0: "GeForce GTX 745"
  CUDA Driver Version / Runtime Version          9.0 / 8.0
  CUDA Capability Major/Minor version number:    5.0
  Total amount of global memory:                 4041 MBytes (4237164544 bytes)
  ( 3) Multiprocessors, (128) CUDA Cores/MP:     384 CUDA Cores
  GPU Max Clock rate:                            1032 MHz (1.03 GHz)
  Memory Clock rate:                             900 Mhz
  Memory Bus Width:                              128-bit
  L2 Cache Size:                                 2097152 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(65536), 2D=(65536, 65536), 3D=(4096, 4096, 4096)
  Maximum Layered 1D Texture Size, (num) layers  1D=(16384), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(16384, 16384), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  2048
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 1 copy engine(s)
  Run time limit on kernels:                     Yes
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 1 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 9.0, CUDA Runtime Version = 8.0, NumDevs = 1, Device0 = GeForce GTX 745
Result = PASS

若结果是​​result=pass​​，就说明成功。

2、查看安装的版本：

bit@bit-613:/$ 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

3、编译cuda例子：

​​javascript:void(0)​​

四、安装cuDNN

4.1 查找对应CUDA版本的cuDNN版本。

​​查找对应CUDA版本的cuDNN版本：​​

• Cuda compilation tools, release 8.0, V8.0.61
• Driver Version: 384.130

cuda8.0.61 对应的cuDNN版本为 7.1.4-7.2.1.

4.2 到官网下载对应的cuDNN软件

​​cuDNN Archive​​

发现有3个对应的Ubuntu版本的cuDNN：

cuDNN v7.1.4 Runtime Library for Ubuntu16.04 (Deb)
cuDNN v7.1.4 Developer Library for Ubuntu16.04 (Deb)
cuDNN v7.1.4 Code Samples and User Guide for Ubuntu16.04 (Deb)

Runtime 和 Developer 版本区别

---

1. developer library 包含了在Ubuntu系统上开发深度学习时所需的cuDNN头文件，如果你不需要开发编译任何深度学习程序，而只是将其用于运行某些深度学习应用，那么只下载“runtime library”就足够了。

最好分别安装这3个文件。

​​Ubuntu系统—CUDA+cuDNN 安装​​

4.3 安装

安装指导：​​官网链接​​​ 安装deb格式的文件，找到相关的安装指南：​​安装指南​​

bit@bit-613:~/下载$ sudo dpkg -i libcudnn7_7.1.4.18-1+cuda8.0_amd64.deb 
[sudo] bit 的密码： 
正在选中未选择的软件包 libcudnn7。
(正在读取数据库 ... 系统当前共安装有 268504 个文件和目录。)
正准备解包 libcudnn7_7.1.4.18-1+cuda8.0_amd64.deb  ...
正在解包 libcudnn7 (7.1.4.18-1+cuda8.0) ...
正在设置 libcudnn7 (7.1.4.18-1+cuda8.0) ...
正在处理用于 libc-bin (2.23-0ubuntu11.2) 的触发器 ...
bit@bit-613:~/下载$ sudo dpkg -i libcudnn7-dev_7.1.4.18-1+cuda8.0_amd64.deb 
正在选中未选择的软件包 libcudnn7-dev。
(正在读取数据库 ... 系统当前共安装有 268511 个文件和目录。)
正准备解包 libcudnn7-dev_7.1.4.18-1+cuda8.0_amd64.deb  ...
正在解包 libcudnn7-dev (7.1.4.18-1+cuda8.0) ...
正在设置 libcudnn7-dev (7.1.4.18-1+cuda8.0) ...
update-alternatives: 使用 /usr/include/x86_64-linux-gnu/cudnn_v7.h 来在自动模式中提供 /usr/include/cudnn.h (libcudnn)
bit@bit-613:~/下载$ sudo dpkg -i libcudnn7-doc_7.1.4.18-1+cuda8.0_amd64.deb 
正在选中未选择的软件包 libcudnn7-doc。
(正在读取数据库 ... 系统当前共安装有 268517 个文件和目录。)
正准备解包 libcudnn7-doc_7.1.4.18-1+cuda8.0_amd64.deb  ...
正在解包 libcudnn7-doc (7.1.4.18-1+cuda8.0) ...
正在设置 libcudnn7-doc (7.1.4.18-1+cuda8.0) ...
bit@bit-613:~/下载$ 
bit@bit-613:~/下载$ 
bit@bit-613:~/下载$ sudo cp cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64.deb 
Anaconda3-5.2.0-Linux-x86_64.sh
cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64.deb
libcudnn7_7.1.4.18-1+cuda8.0_amd64.deb
libcudnn7-dev_7.1.4.18-1+cuda8.0_amd64.deb
libcudnn7-doc_7.1.4.18-1+cuda8.0_amd64.deb
sogoupinyin_2.3.1.0112_amd64.deb
未确认 514476.crdownload
bit@bit-613:~/下载$ ls

4.4 测试

cp -r /usr/src/cudnn_samples_v7/  /home/bit/
cd  /home/bit/cudnn_samples_v7/mnistCUDNN
make clean && make
./mnistCUDNN

发现输出​​Test passed!​​后，说明成功：

bit@bit-613:~/cudnn_samples_v7/mnistCUDNN$ make clean && make
nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
rm -rf *o
rm -rf mnistCUDNN
nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
/usr/local/cuda/bin/nvcc -ccbin g++ -I/usr/local/cuda/include -IFreeImage/include  -m64    -gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_53,code=sm_53 -gencode arch=compute_53,code=compute_53 -o fp16_dev.o -c fp16_dev.cu
g++ -I/usr/local/cuda/include -IFreeImage/include   -o fp16_emu.o -c fp16_emu.cpp
g++ -I/usr/local/cuda/include -IFreeImage/include   -o mnistCUDNN.o -c mnistCUDNN.cpp
/usr/local/cuda/bin/nvcc -ccbin g++   -m64      -gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_53,code=sm_53 -gencode arch=compute_53,code=compute_53 -o mnistCUDNN fp16_dev.o fp16_emu.o mnistCUDNN.o  -LFreeImage/lib/linux/x86_64 -LFreeImage/lib/linux -lcudart -lcublas -lcudnn -lfreeimage -lstdc++ -lm
bit@bit-613:~/cudnn_samples_v7/mnistCUDNN$ 
bit@bit-613:~/cudnn_samples_v7/mnistCUDNN$ ./mnistCUDNN 
cudnnGetVersion() : 7104 , CUDNN_VERSION from cudnn.h : 7104 (7.1.4)
Host compiler version : GCC 5.4.0
There are 1 CUDA capable devices on your machine :
device 0 : sms  3  Capabilities 5.0, SmClock 1032.5 Mhz, MemSize (Mb) 4040, MemClock 900.0 Mhz, Ecc=0, boardGroupID=0
Using device 0

Testing single precision
Loading image data/one_28x28.pgm
Performing forward propagation ...
Testing cudnnGetConvolutionForwardAlgorithm ...
Fastest algorithm is Algo 1
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: 0.036864 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: 0.039200 time requiring 3464 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: 0.058112 time requiring 57600 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: 0.211584 time requiring 207360 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: 0.559648 time requiring 2057744 memory
Resulting weights from Softmax:
0.0000000 0.9999399 0.0000000 0.0000000 0.0000561 0.0000000 0.0000012 0.0000017 0.0000010 0.0000000 
Loading image data/three_28x28.pgm
Performing forward propagation ...
Resulting weights from Softmax:
0.0000000 0.0000000 0.0000000 0.9999288 0.0000000 0.0000711 0.0000000 0.0000000 0.0000000 0.0000000 
Loading image data/five_28x28.pgm
Performing forward propagation ...
Resulting weights from Softmax:
0.0000000 0.0000008 0.0000000 0.0000002 0.0000000 0.9999820 0.0000154 0.0000000 0.0000012 0.0000006 

Result of classification: 1 3 5

Test passed!

Testing half precision (math in single precision)
Loading image data/one_28x28.pgm
Performing forward propagation ...
Testing cudnnGetConvolutionForwardAlgorithm ...
Fastest algorithm is Algo 1
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: 0.032224 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: 0.043744 time requiring 3464 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: 0.082944 time requiring 28800 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: 0.218880 time requiring 207360 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: 0.560992 time requiring 2057744 memory
Resulting weights from Softmax:
0.0000001 1.0000000 0.0000001 0.0000000 0.0000563 0.0000001 0.0000012 0.0000017 0.0000010 0.0000001 
Loading image data/three_28x28.pgm
Performing forward propagation ...
Resulting weights from Softmax:
0.0000000 0.0000000 0.0000000 1.0000000 0.0000000 0.0000714 0.0000000 0.0000000 0.0000000 0.0000000 
Loading image data/five_28x28.pgm
Performing forward propagation ...
Resulting weights from Softmax:
0.0000000 0.0000008 0.0000000 0.0000002 0.0000000 1.0000000 0.0000154 0.0000000 0.0000012 0.0000006 

Result of classification: 1 3 5

Test passed!

五、问题

5.1 nvcc warning警告

nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).

原因是，Makefile中采用了CUDA的compute capability 2.0和2.1，这是两种计算能力。安装的CUDA版本是8.0，但 从CUDA 8.0开始compute capability 2.0和2.1被弃用了，所以可以将-gencode arch=compute_20,code=sm_20 和-gencode arch=compute_20,code=sm_21这两行删除即可。

​​nvcc编译器警告’compute_20’………。​​

5.2 无法获取 dpkg 前端锁

bit@bit-613:~/tmp/NVIDIA_CUDA-8.0_Samples$ sudo apt install cmake
E: 无法获得锁 /var/lib/dpkg/lock-frontend - open (11: 资源暂时不可用)
E: 无法获取 dpkg 前端锁 (/var/lib/dpkg/lock-frontend)，是否有其他进程正占用它？

解决:

bit@bit-613:~/tmp/NVIDIA_CUDA-8.0_Samples$ ps -e|grep apt
 1448 ?        00:00:00 apt.systemd.dai
 1700 ?        00:00:00 apt.systemd.dai
bit@bit-613:~/tmp/NVIDIA_CUDA-8.0_Samples$ sudo kill 1448
bit@bit-613:~/tmp/NVIDIA_CUDA-8.0_Samples$ sudo kill 1700

5.3 No MPI compiler found

-------------------------------------------------------------------------------------
WARNING - No MPI compiler found.
-------------------------------------------------------------------------------------
CUDA Sample "simpleMPI" cannot be built without an MPI Compiler.
This will be a dry-run of the Makefile.
For more information on how to set up your environment to build and run this 
sample, please refer the CUDA Samples documentation and release notes
-------------------------------------------------------------------------------------

ubuntu 16.04 mpi安装：​​ mpi​​​、​​ubuntu 虚拟机下配置MPI​​

sudo apt-get install gfortran
wget http://www.mpich.org/static/downloads/3.3/mpich-3.3.tar.gz

sudo tar -zxvf mpich-3.2.tar.gz
cd mpich-3.3
./configure
make
make install

5.4 cudnn.h: 没有那个文件或目录

cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2
cat: /usr/local/cuda/include/cudnn.h: 没有那个文件或目录

1、找cudnn.h文件

find / -name cudnn.h

• 1

这个命令会列出所有的cudnn.h文件
找到自己安装cuda时的虚拟环境中的cudnn.h，我的是：​​​/usr/include/cudnn.h​​​

2、将找到的cudnn.h文件复制到

cp /usr/include/cudnn.h /usr/local/cuda/include/

3、查询cudnn的版本：

root@bit-613:~# cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2
#define CUDNN_MAJOR 7
#define CUDNN_MINOR 1
#define CUDNN_PATCHLEVEL 4
--
#define CUDNN_VERSION    (CUDNN_MAJOR * 1000 + CUDNN_MINOR * 100 + CUDNN_PATCHLEVEL)

#include "driver_types.h"

4、 查询

find / -name libcudnn.so

文件在：​​/usr/lib/x86_64-linux-gnu/libcudnn.so​​

复制到：

cp /usr/lib/x86_64-linux-gnu/libcudnn.so   /usr/local/cuda-8.0/lib64

然后make，生成了可执行文件darknet和libdarknet.so及库libdarknet.a
大功告成。