Step1. 准备工作

sudo apt install vim

sudo apt-get update

Step2. 安装CUDA9.1

[1]

lspci | grep -i nvidia

01:00.0 VGA compatible controller: NVIDIA Corporation Device 1b81 (rev a1)

01:00.1 Audio device: NVIDIA Corporation Device 10f0 (rev a1)

[2]

uname -m && cat /etc/*release

x86_64

DISTRIB_ID=Ubuntu

DISTRIB_RELEASE=16.04

DISTRIB_CODENAME=xenial

DISTRIB_DESCRIPTION="Ubuntu 16.04.3 LTS"

NAME="Ubuntu"

VERSION="16.04.3 LTS (Xenial Xerus)"

ID=ubuntu

ID_LIKE=debian

PRETTY_NAME="Ubuntu 16.04.3 LTS"

VERSION_ID="16.04"

HOME_URL=""

SUPPORT_URL=""

BUG_REPORT_URL=""

VERSION_CODENAME=xenial

UBUNTU_CODENAME=xenial

[3]

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.

[4]

sudo apt-get install linux-headers-$(uname -r)

[sudo] sanray 的密码：

正在读取软件包列表... 完成

正在分析软件包的依赖关系树

正在读取状态信息... 完成

linux-headers-4.13.0-26-generic 已经是最新版 (4.13.0-26.29~16.04.2)。

linux-headers-4.13.0-26-generic 已设置为手动安装。

下列软件包是自动安装的并且现在不需要了：

linux-headers-4.10.0-28 linux-headers-4.10.0-28-generic linux-image-4.10.0-28-generic

linux-image-extra-4.10.0-28-generic

使用'sudo apt autoremove'来卸载它(它们)。

升级了 0 个软件包，新安装了 0 个软件包，要卸载 0 个软件包，有 180 个软件包未被升级。

[5]

修改软件下载源（这一步非常重要，国外网站很可能无法下载或者速度非常慢）

[6]

下载安装

重启reboot，系统加载NVIDIA drivers

[7]

确认路径存在

ls /usr/local/cuda-9.1/bin/

设置环境路径

export PATH=/usr/local/cuda-9.1/bin${PATH:+:${PATH}}

[8]

cd /usr/lib/systemd/

sudo mkdir system

cd system/

创建/usr/lib/systemd/system/nvidia-persistenced.service文件

sudo vi nvidia-persistenced.service

[Unit]

Description=NVIDIA Persistence Daemon

Wants=syslog.target

[Service]

Type=forking

PIDFile=/var/run/nvidia-persistenced/nvidia-persistenced.pid

Restart=always

ExecStart=/usr/bin/nvidia-persistenced --verbose

ExecStopPost=/bin/rm -rf /var/run/nvidia-persistenced

[Install]

WantedBy=multi-user.target

运行

sudo systemctl enable nvidia-persistenced

[9]

修改rules文件

ls /lib/udev/rules.d/40-vm-hotadd.rules

sudo vim /lib/udev/rules.d/40-vm-hotadd.rules

修改前

# Memory hotadd request

SUBSYSTEM=="memory", ACTION=="add", DEVPATH=="/devices/system/memory/memory[0-9]*", TEST=="state", ATTR{state}="online"

修改后

# Memory hotadd request

# 2018-01-18 be commendted out by wanghonggang(18576671185), for the NVIDIA

# CUDA driver to function properly on POWER9 systems.

#SUBSYSTEM=="memory", ACTION=="add", DEVPATH=="/devices/system/memory/memory[0-9]*", TEST=="state", ATTR{state}="online"

================================================

验证结果

cat /proc/driver/nvidia/version

NVRM version: NVIDIA UNIX x86_64 Kernel Module 387.34 Tue Nov 21 03:09:00 PST 2017

GCC version: gcc version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.5)

nvcc -V

nvcc: NVIDIA (R) Cuda compiler driver

Copyright (c) 2005-2017 NVIDIA Corporation

Built on Fri_Nov__3_21:07:56_CDT_2017

Cuda compilation tools, release 9.1, V9.1.85

安装第三方库

sudo apt-get install g++ freeglut3-dev build-essential libx11-dev \

libxmu-dev libxi-dev libglu1-mesa libglu1-mesa-dev

安装samples

cuda-install-samples-9.1.sh <dir>

ls /usr/local/cuda-9.1/samples/

cd ~/NVIDIA_CUDA-9.1_Samples

make

生成bin文件夹和很多文件

cd ~/NVIDIA_CUDA-9.1_Samples/bin/x86_64/linux/release

检查结果

[错误1]

/usr/local/cuda-9.1/samples/bin/x86_64/linux/release/deviceQuery

/usr/local/cuda-9.1/samples/bin/x86_64/linux/release/deviceQuery Starting...

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

cudaGetDeviceCount returned 35

-> CUDA driver version is insufficient for CUDA runtime version

Result = FAIL

[错误2]

/usr/local/cuda-9.1/samples/bin/x86_64/linux/release/deviceQuery

/usr/local/cuda-9.1/samples/bin/x86_64/linux/release/deviceQuery Starting...

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

cudaGetDeviceCount returned 38

-> no CUDA-capable device is detected

Result = FAIL

[正确]

/usr/local/cuda-9.1/samples/bin/x86_64/linux/release/deviceQuery

/usr/local/cuda-9.1/samples/bin/x86_64/linux/release/deviceQuery Starting...

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

Detected 1 CUDA Capable device(s)

Device 0: "GeForce GTX 1070"

CUDA Driver Version / Runtime Version 9.1 / 9.1

CUDA Capability Major/Minor version number: 6.1

Total amount of global memory: 8106 MBytes (8499691520 bytes)

(15) Multiprocessors, (128) CUDA Cores/MP: 1920 CUDA Cores

GPU Max Clock rate: 1721 MHz (1.72 GHz)

Memory Clock rate: 4004 Mhz

Memory Bus Width: 256-bit

L2 Cache Size: 2097152 bytes

Maximum Texture Dimension Size (x,y,z) 1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)

Maximum Layered 1D Texture Size, (num) layers 1D=(32768), 2048 layers

Maximum Layered 2D Texture Size, (num) layers 2D=(32768, 32768), 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 2 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

Supports Cooperative Kernel Launch: Yes

Supports MultiDevice Co-op Kernel Launch: 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 = 9.1, CUDA Runtime Version = 9.1, NumDevs = 1

Result = PASS

~/NVIDIA_CUDA-9.1_Samples/bin/x86_64/linux/release$ ./bandwidthTest

[CUDA Bandwidth Test] - Starting...

Running on...

Device 0: GeForce GTX 1070

Quick Mode

Host to Device Bandwidth, 1 Device(s)

PINNED Memory Transfers

Transfer Size (Bytes)Bandwidth(MB/s)

3355443212374.0

Device to Host Bandwidth, 1 Device(s)

PINNED Memory Transfers

Transfer Size (Bytes)Bandwidth(MB/s)

3355443212351.9

Device to Device Bandwidth, 1 Device(s)

PINNED Memory Transfers

Transfer Size (Bytes)Bandwidth(MB/s)

33554432191159.3

Result = PASS

NOTE: The CUDA Samples are not meant for performance measurements. Results may vary when GPU Boost is enabled.

大量文档在这里

/usr/local/cuda-9.1/doc