Ubuntu16.04+GTX2070+Driver418.43+CUDA10.1+cuDNN7.6

最近需要用到一台服务器的GPU跑实验，其间 COLMAP 编译过程出错，提示 cuda 版本不支持，cmake虽然通过了，但其实没有找到支持的CUDA架构。

cv@cv:~/mvs_project/colmap/build$ cmake ..
...
-- Automatic GPU detection failed. Building for common architectures.
-- Autodetected CUDA architecture(s): 3.0;3.5;5.0;5.2;6.0;6.1;7.0;7.0+PTX
-- Enabling CUDA support (version: 9.0, archs: sm_30 sm_35 sm_50 sm_52 sm_60 sm_61 sm_70 compute_70)
...

cv@cv:~/mvs_project/colmap/build$ make
[  %] Automatic rcc for target flann
[  %] Built target flann_automoc
[  %] Building CXX object lib/FLANN/CMakeFiles/flann.dir/flann.cpp.o
[  %] Building C object lib/FLANN/CMakeFiles/flann.dir/ext/lz4.c.o
[  %] Building C object lib/FLANN/CMakeFiles/flann.dir/ext/lz4hc.c.o
[  %] Linking CXX static library libflann.a
[  %] Built target flann
[  %] Automatic rcc for target graclus
[  %] Built target graclus_automoc
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/util.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mincover.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kwayrefine.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/refine.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/ometis.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mmatch.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mutil.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mpmetis.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/balance.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mfm2.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mesh.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/compress.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/initpart.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/subdomains.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kwayvolfm.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/fortran.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/pmetis.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kwayfm.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/parmetis.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/coarsen.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mkwayfmh.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mbalance2.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mbalance.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mmd.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/pqueue.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/estmem.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/myqsort.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kvmetis.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/fm.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/ccgraph.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/minitpart2.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/bucketsort.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/graph.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mrefine2.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/frename.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/stat.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/debug.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/srefine.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/meshpart.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mrefine.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kwayvolrefine.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/match.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kmetis.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mkwayrefine.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/metis.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mcoarsen.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/timing.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mfm.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/memory.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/minitpart.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/sfm.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mkmetis.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/separator.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/multilevelLib/wkkm.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/multilevelLib/mlkkm.c.o
[ %] Linking C static library libgraclus.a
[ %] Built target graclus
[ %] Automatic rcc for target lsd
[ %] Built target lsd_automoc
[ %] Building C object lib/LSD/CMakeFiles/lsd.dir/lsd.c.o
[ %] Linking C static library liblsd.a
[ %] Built target lsd
[ %] Automatic rcc for target pba
[ %] Built target pba_automoc
[ %] Building NVCC (Device) object lib/PBA/CMakeFiles/pba.dir/pba_generated_ProgramCU.cu.o
CMake Error at pba_generated_ProgramCU.cu.o.cmake: (message):
  Error generating
  /home/cv/mvs_project/colmap/build/lib/PBA/CMakeFiles/pba.dir//./pba_generated_ProgramCU.cu.o
 
lib/PBA/CMakeFiles/pba.dir/build.make:: recipe for target 'lib/PBA/CMakeFiles/pba.dir/pba_generated_ProgramCU.cu.o' failed
make[]: *** [lib/PBA/CMakeFiles/pba.dir/pba_generated_ProgramCU.cu.o] Error
CMakeFiles/Makefile2:: recipe for target 'lib/PBA/CMakeFiles/pba.dir/all' failed
make[]: *** [lib/PBA/CMakeFiles/pba.dir/all] Error
Makefile:: recipe for target 'all' failed
make: *** [all] Error

colmap_build_error

于是又开始配置环境，首先根据自己机器配置NVIDIA官方网站下载 GeForce 驱动程序

>> 检查机器环境及配置

内核版本及操作系统信息

cv@cv:~/mvs_project/colmap/build$ uname -r
4.15.0-65-generic
 
cv@cv:~/mvs_project/colmap/build$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description:    Ubuntu 16.04. LTS
Release:        16.04
Codename:       xenial
 
cv@cv:~/mvs_project/colmap/build$ gcc --version
gcc (Ubuntu 5.4.-6ubuntu1~16.04.) 5.4.
Copyright (C)  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.

已经安装过显卡驱动的机器可以直接通过 nvidia-smi 命令显示显卡型号和驱动版本信息

cv@cv:~/mvs_project/colmap/build$ nvidia-smi
Sat Nov  ::
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 418.43       Driver Version: 418.43       CUDA Version: 10.1     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|     GeForce RTX     Off  | ::00.0 Off |                  N/A |
|  %   65C    P0     1W / 210W |      0MiB /  7952MiB |      %      Default |
+-------------------------------+----------------------+----------------------+
 
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

对尚未安装过显卡驱动的机器，可以通过 lspci 指令查询，grep -i 的意思是忽略后面匹配项的大小写

cv@cv:~/mvs_project/colmap/build$ lspci | grep -i vga | grep -i nvidia
:00.0 VGA compatible controller: NVIDIA Corporation Device 1f07 (rev a1)

这里返回的是一串十六进制代码 1f07，跟我们平常所见略有不同，需要翻译一下，到 PCI devices 查询。打不开网页或者打开很慢的可以参考放在GitHub上的一份常见型号对应表

知道了自己的机器的配置就可以到上面给出的网站(https://www.geforce.cn/drivers)下载对应的驱动程序。

开始安装驱动之前的准备工作

>> 卸载旧版本或安装失败的驱动
cv@cv:~/mvs_project/colmap/build$ cd
cv@cv:~$ sudo ./NVIDIA-Linux-x86_64-418.43.run --uninstall
>> 安装可能需要的依赖
cv@cv:~$ sudo apt update
cv@cv:~$ sudo apt install dkms build-essential linux-headers-generic
cv@cv:~$ sudo apt install gcc-multilib xorg-dev
cv@cv:~$ sudo apt install freeglut3-dev libx11-dev libxmu-dev libxi-dev
cv@cv:~$ sudo apt install libgl1-mesa-glx libglu1-mesa libglu1-mesa-dev
>> 禁用 NOUVEAU 驱动

直接使用 VIM 打开，没有该文件时自动新建
cv@cv:~$ sudo vim /etc/modprobe.d/blacklist-nouveau.conf
在文件中添加如下内容，保存退出
blacklist nouveau
blacklist lbm-nouveau
options nouveau modeset=
alias nouveau off
alias lbm-nouveau off
然后执行下面的指令，禁用 nouveau 内核模块，更新配置，重启
cv@cv:~$ echo options nouveau modeset= | sudo tee -a /etc/modprobe.d/nouveau-kms.conf
cv@cv:~$ sudo update-initramfs -u
cv@cv:~$ sudo reboot
CTRL+ALT+F1 进入命令行模式，输入下面的命令，如果没有任何显示则表明禁用驱动成功了。然后关闭图形界面，后面要记得重新打开。
cv@cv:~$ lsmod | grep nouveau
cv@cv:~$ sudo service lightdm stop

然后开始安装显卡驱动

cv@cv:~$ chmod a+x NVIDIA-Linux-x86_64-418.43.run
cv@cv:~$ sudo ./NVIDIA-Linux-x86_64-418.43.run --dkms --no-opengl-files
-dkms　　默认开启。在 kernel 自行更新时将驱动程序安装至模块中，从而阻止驱动程序重新安装。

–no-opengl-files　　表示只安装驱动文件，不安装OpenGL文件。这个参数不可省略，否则会导致登陆界面死循环。因为NVIDIA的驱动默认会安装OpenGL，而Ubuntu的内核本身也有OpenGL且与GUI显示息息相关，

一旦NVIDIA的驱动覆盖了OpenGL，在GUI需要动态链接OpenGL库的时候就会出现问题。

–no-x-check　　表示安装驱动时不检查X服务，非必需，已经禁用图形界面。

–no-nouveau-check　　表示安装驱动时不检查nouveau，非必需，已经禁用nouveau驱动。

–disable-nouveau　　禁用nouveau。非必需，因为之前已经手动禁用了nouveau。

安装过程中弹出pre-install script failed的信息，继续安装即可，没有影响。

dkms 选项选yes

32位兼容选项选yes

x-org 选项保持默认选no

安装完成后打开图形桌面。
cv@cv:~$ sudo service lightdm start
cv@cv:~$ nvidia-smi
如果有显示GPU相关信息表示驱动安装成功。

卸载CUDA

首先卸载以前安装的或安装失败的CUDA，以便我们顺利进行下面的步骤，直接执行CUDA自带的卸载脚本。
cv@cv:~$ sudo /usr/local/cuda-9.0/bin/uninstall_cuda_9..pl
卸载完成后，清除残留文件夹。
cv@cv:~$ sudo rm -rf /usr/local/cuda-9.0/

安装CUDA和CUDNN

>> 首先下载安装文件，我们要安装的是CUDA10.1和CUDNN7.6

根据对应关系到 CUDA 下载页面寻找自己需要的版本，比如我下载的是 CUDA Toolkit 10.1 update2，选择好操作系统，系统架构和安装类型之后下载即可。

到 CUDA Toolkit Archive 网站上下载

cuda_10.1.243_418.87.00_linux.run

然后下载CUDNN，需要注册或登录NVIDIA账号，看清楚版本，到 cuDNN Download 网站上 for CUDA 10.1 下载里面的三个deb安装包

libcudnn7_7.6.5.32-1+cuda10.1_amd64.deb

libcudnn7-dev_7.6.5.32-1+cuda10.1_amd64.deb

libcudnn7-doc_7.6.5.32-1+cuda10.1_amd64.deb

>> 然后开始安装 CUDA
cv@cv:~$ sudo service lightdm stop
cv@cv:~$ chmod a+x cuda_10..243_418..00_linux.run
cv@cv:~$ sudo ./cuda_10..243_418..00_linux.run
是否同意条款 accept

选择安装界面，除了418.87取消勾选之外其他保持默认

剩下的都保持默认即可

然后打开配置文件，并在末尾添加链接路径，保存退出
cv@cv:~$ vim ~/.bashrc
export PATH=/usr/local/cuda-10.1/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
使生效
cv@cv:~$ source ~/.bashrc
这是应该已经可以查看CUDA安装版本了
cv@cv:~$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 20xx-20xx NVIDIA Corporation
Built on Tue_Jan_10_13::03_CST_20xx
Cuda compilation tools, release 10.1, V10..x
>> 接着安装 cuDNN
cv@cv:~$ sudo dpkg -i libcudnn7_7.6.5.-+cuda10.1_amd64.deb
cv@cv:~$ sudo dpkg -i libcudnn7-dev_7.6.5.-+cuda10.1_amd64.deb
cv@cv:~$ sudo dpkg -i libcudnn7-doc_7.6.5.-+cuda10.1_amd64.deb
>> 打开图形界面
cv@cv:~$ sudo service lightdm start

验证安装是否成功

>> CUDA 测试，进入到 CUDA 例程路径下，编译并测试

cv@cv:~$ cd NVIDIA_CUDA-.1_Samples/
cv@cv:~/NVIDIA_CUDA-.1_Samples$ make
cv@cv:~/NVIDIA_CUDA-.1_Samples$ cd bin/x86_64/linux/release/

cv@cv:~/NVIDIA_CUDA-.1_Samples/bin/x86_64/linux/release$ ./deviceQuery
./deviceQuery Starting...
 
 CUDA Device Query (Runtime API) version (CUDART static linking)
 
Detected  CUDA Capable device(s)
 
Device : "GeForce RTX 2070"
  CUDA Driver Version / Runtime Version          10.1 / 10.1
  CUDA Capability Major/Minor version number:    7.5
  Total amount of global memory:                  MBytes ( bytes)
  () Multiprocessors, ( ) CUDA Cores/MP:      CUDA Cores
  GPU Max Clock rate:                             MHz (1.71 GHz)
  Memory Clock rate:                              Mhz
  Memory Bus Width:                              -bit
  L2 Cache Size:                                  bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(), 2D=(, ), 3D=(, , )
  Maximum Layered 1D Texture Size, (num) layers  1D=(),  layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(, ),  layers
  Total amount of constant memory:                bytes
  Total amount of shared memory per block:        bytes
  Total number of registers available per block:
  Warp size:
  Maximum number of threads per multiprocessor:
  Maximum number of threads per block:
  Max dimension size of a thread block (x,y,z): (, , )
  Max dimension size of a grid size    (x,y,z): (, , )
  Maximum memory pitch:                           bytes
  Texture alignment:                              bytes
  Concurrent copy and kernel execution:          Yes with  copy engine(s)
  Run time limit on kernels:                     No
  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 supports Compute Preemption:            Yes
  Supports Cooperative Kernel Launch:            Yes
  Supports MultiDevice Co-op Kernel Launch:      Yes
  Device PCI Domain ID / Bus ID / location ID:    /  /
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >
 
deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 10.1, CUDA Runtime Version = 10.1, NumDevs =
Result = PASS

cv@cv:~/NVIDIA_CUDA-.1_Samples/bin/x86_64/linux/release$ ./bandwidthTest
[CUDA Bandwidth Test] - Starting...
Running on...
 
 Device : GeForce RTX
 Quick Mode
 
 Host to Device Bandwidth,  Device(s)
 PINNED Memory Transfers
   Transfer Size (Bytes)        Bandwidth(GB/s)
                        12.8
 
 Device to Host Bandwidth,  Device(s)
 PINNED Memory Transfers
   Transfer Size (Bytes)        Bandwidth(GB/s)
                        13.1
 
 Device to Device Bandwidth,  Device(s)
 PINNED Memory Transfers
   Transfer Size (Bytes)        Bandwidth(GB/s)
                        382.0
 
Result = PASS
 
NOTE: The CUDA Samples are not meant for performance measurements. Results may vary when GPU Boost is enabled.

>> cuDNN 测试

cv@cv:~$ cat /usr/include/cudnn.h | grep CUDNN_MAJOR -A  -m
#define CUDNN_MAJOR 7
#define CUDNN_MINOR 6
#define CUDNN_PATCHLEVEL 5

cv@cv:~$ cp -r /usr/src/cudnn_samples_v7/ .
cv@cv:~$ cd cudnn_samples_v7/mnistCUDNN/
cv@cv:~$ make
Linking agains cublasLt = true
CUDA VERSION:
TARGET ARCH: x86_64
HOST_ARCH: x86_64
TARGET OS: linux
SMS:
/usr/local/cuda/bin/nvcc -ccbin g++ -I/usr/local/cuda/include -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_60,code=sm_60 -gencode arch=compute_61,code=sm_61
-gencode arch=compute_62,code=sm_62 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_72,code=sm_72
-gencode arch=compute_75,code=sm_75 -gencode arch=compute_75,code=compute_75 -o fp16_dev.o -c fp16_dev.cu
g++ -I/usr/local/cuda/include -I/usr/local/cuda/include -IFreeImage/include   -o fp16_emu.o -c fp16_emu.cpp
g++ -I/usr/local/cuda/include -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_60,code=sm_60 -gencode arch=compute_61,code=sm_61
-gencode arch=compute_62,code=sm_62 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_72,code=sm_72
-gencode arch=compute_75,code=sm_75 -gencode arch=compute_75,code=compute_75 -o mnistCUDNN fp16_dev.o fp16_emu.o mnistCUDNN.o
-I/usr/local/cuda/include -I/usr/local/cuda/include -IFreeImage/include
-L/usr/local/cuda/lib64 -L/usr/local/cuda/lib64 -lcublasLt -LFreeImage/lib/linux/x86_64
-LFreeImage/lib/linux -lcudart -lcublas -lcudnn -lfreeimage -lstdc++ -lm

cv@cv:~/cudnn_samples_v7/mnistCUDNN$ ./mnistCUDNN
cudnnGetVersion() :  , CUDNN_VERSION from cudnn.h :  (7.6.)
Host compiler version : GCC 5.4.
There are  CUDA capable devices on your machine :
device  : sms   Capabilities 7.5, SmClock 1710.0 Mhz, MemSize (Mb) , MemClock 7001.0 Mhz, Ecc=, boardGroupID=
Using device 
 
Testing single precision
Loading image data/one_28x28.pgm
Performing forward propagation ...
Testing cudnnGetConvolutionForwardAlgorithm ...
Fastest algorithm is Algo
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo : 0.039040 time requiring  memory
^^^^ CUDNN_STATUS_SUCCESS for Algo : 0.100576 time requiring  memory
^^^^ CUDNN_STATUS_SUCCESS for Algo : 0.122400 time requiring  memory
^^^^ CUDNN_STATUS_SUCCESS for Algo : 0.130560 time requiring  memory
^^^^ CUDNN_STATUS_SUCCESS for Algo : 0.173888 time requiring  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:   
 
Test passed!
 
Testing half precision (math in single precision)
Loading image data/one_28x28.pgm
Performing forward propagation ...
Testing cudnnGetConvolutionForwardAlgorithm ...
Fastest algorithm is Algo
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo : 0.022528 time requiring  memory
^^^^ CUDNN_STATUS_SUCCESS for Algo : 0.061344 time requiring  memory
^^^^ CUDNN_STATUS_SUCCESS for Algo : 0.065536 time requiring  memory
^^^^ CUDNN_STATUS_SUCCESS for Algo : 0.070208 time requiring  memory
^^^^ CUDNN_STATUS_SUCCESS for Algo : 0.082592 time requiring  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:   
 
Test passed!

当这些配置好之后，COLMAP 的编译就很顺利地通过了。

cv@cv:~/mvs_project/colmap/build$ cmake ..
-- The C compiler identification is GNU 5.4.
-- The CXX compiler identification is GNU 5.4.
-- 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
-- Found installed version of Eigen: /usr/lib/cmake/eigen3
-- Found required Ceres dependency: Eigen version 3.2. in /usr/include/eigen3
-- Found required Ceres dependency: glog
-- Performing Test GFLAGS_IN_GOOGLE_NAMESPACE
-- Performing Test GFLAGS_IN_GOOGLE_NAMESPACE - Success
-- Found required Ceres dependency: gflags
-- Found Ceres version: 1.14. installed in: /usr/local with components: [EigenSparse, SparseLinearAlgebraLibrary, LAPACK, SuiteSparse, CXSparse, SchurSpecializations, OpenMP, Multithreading]
-- Boost version: 1.58.
-- Found the following Boost libraries:
--   program_options
--   filesystem
--   graph
--   regex
--   system
--   unit_test_framework
-- Found Eigen3: /usr/include/eigen3 (Required is at least version "2.91.0")
-- Found Eigen
--   Includes : /usr/include/eigen3
-- Found FreeImage
--   Includes : /usr/include
--   Libraries : /usr/lib/x86_64-linux-gnu/libfreeimage.so
-- Found Glog
--   Includes : /usr/include
--   Libraries : /usr/lib/x86_64-linux-gnu/libglog.so
-- Found OpenGL: /usr/lib/x86_64-linux-gnu/libGL.so
-- Found Glew
--   Includes : /usr/include
--   Libraries : /usr/lib/x86_64-linux-gnu/libGLEW.so
-- Found Git: /usr/bin/git (found version "2.7.4")
-- Found Threads: TRUE
-- Found Qt
--   Module : /usr/lib/x86_64-linux-gnu/cmake/Qt5Core
--   Module : /usr/lib/x86_64-linux-gnu/cmake/Qt5OpenGL
--   Module : /usr/lib/x86_64-linux-gnu/cmake/Qt5Widgets
-- Found CGAL
--   Includes : /usr/include
--   Libraries : /usr/lib/x86_64-linux-gnu/libCGAL.so.11.0.
-- Build type not specified, using Release
-- Enabling SIMD support
-- Enabling OpenMP support
-- Disabling interprocedural optimization
-- Autodetected CUDA architecture(s):  7.5
-- Enabling CUDA support (version: 10.1, archs: sm_75)
-- Enabling OpenGL support
-- Disabling profiling support
-- Enabling CGAL support
-- Configuring done
-- Generating done
-- Build files have been written to: /home/cv/mvs_project/colmap/build
 
cv@cv:~/mvs_project/colmap/build$ make
Scanning dependencies of target flann_automoc
[  %] Automatic rcc for target flann
[  %] Built target flann_automoc
Scanning dependencies of target flann
[  %] Building CXX object lib/FLANN/CMakeFiles/flann.dir/flann.cpp.o
[  %] Building C object lib/FLANN/CMakeFiles/flann.dir/ext/lz4.c.o
[  %] Building C object lib/FLANN/CMakeFiles/flann.dir/ext/lz4hc.c.o
[  %] Linking CXX static library libflann.a
[  %] Built target flann
Scanning dependencies of target graclus_automoc
[  %] Automatic rcc for target graclus
[  %] Built target graclus_automoc
Scanning dependencies of target graclus
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/util.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mincover.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kwayrefine.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/refine.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/ometis.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mmatch.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mutil.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mpmetis.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/balance.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mfm2.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mesh.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/compress.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/initpart.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/subdomains.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kwayvolfm.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/fortran.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/pmetis.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kwayfm.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/parmetis.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/coarsen.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mkwayfmh.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mbalance2.c.o
[  %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mbalance.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mmd.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/pqueue.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/estmem.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/myqsort.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kvmetis.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/fm.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/ccgraph.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/minitpart2.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/bucketsort.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/graph.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mrefine2.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/frename.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/stat.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/debug.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/srefine.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/meshpart.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mrefine.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kwayvolrefine.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/match.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/kmetis.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mkwayrefine.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/metis.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mcoarsen.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/timing.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mfm.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/memory.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/minitpart.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/sfm.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/mkmetis.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/metisLib/separator.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/multilevelLib/wkkm.c.o
[ %] Building C object lib/Graclus/CMakeFiles/graclus.dir/multilevelLib/mlkkm.c.o
[ %] Linking C static library libgraclus.a
[ %] Built target graclus
Scanning dependencies of target lsd_automoc
[ %] Automatic rcc for target lsd
[ %] Built target lsd_automoc
Scanning dependencies of target lsd
[ %] Building C object lib/LSD/CMakeFiles/lsd.dir/lsd.c.o
[ %] Linking C static library liblsd.a
[ %] Built target lsd
Scanning dependencies of target pba_automoc
[ %] Automatic rcc for target pba
[ %] Built target pba_automoc
[ %] Building NVCC (Device) object lib/PBA/CMakeFiles/pba.dir/pba_generated_ProgramCU.cu.o
Scanning dependencies of target pba
[ %] Building CXX object lib/PBA/CMakeFiles/pba.dir/ConfigBA.cpp.o
[ %] Building CXX object lib/PBA/CMakeFiles/pba.dir/CuTexImage.cpp.o
[ %] Building CXX object lib/PBA/CMakeFiles/pba.dir/pba.cpp.o
[ %] Building CXX object lib/PBA/CMakeFiles/pba.dir/SparseBundleCPU.cpp.o
[ %] Building CXX object lib/PBA/CMakeFiles/pba.dir/SparseBundleCU.cpp.o
[ %] Linking CXX static library libpba.a
[ %] Built target pba
Scanning dependencies of target poisson_recon_automoc
[ %] Automatic rcc for target poisson_recon
[ %] Built target poisson_recon_automoc
Scanning dependencies of target poisson_recon
[ %] Building CXX object lib/PoissonRecon/CMakeFiles/poisson_recon.dir/CmdLineParser.cpp.o
[ %] Building CXX object lib/PoissonRecon/CMakeFiles/poisson_recon.dir/Factor.cpp.o
[ %] Building CXX object lib/PoissonRecon/CMakeFiles/poisson_recon.dir/Geometry.cpp.o
[ %] Building CXX object lib/PoissonRecon/CMakeFiles/poisson_recon.dir/MarchingCubes.cpp.o
[ %] Building CXX object lib/PoissonRecon/CMakeFiles/poisson_recon.dir/PlyFile.cpp.o
[ %] Building CXX object lib/PoissonRecon/CMakeFiles/poisson_recon.dir/PoissonRecon.cpp.o
[ %] Building CXX object lib/PoissonRecon/CMakeFiles/poisson_recon.dir/SurfaceTrimmer.cpp.o
[ %] Linking CXX static library libpoisson_recon.a
[ %] Built target poisson_recon
Scanning dependencies of target sift_gpu_automoc
[ %] Automatic rcc for target sift_gpu
[ %] Built target sift_gpu_automoc
[ %] Building NVCC (Device) object lib/SiftGPU/CMakeFiles/sift_gpu.dir/sift_gpu_generated_ProgramCU.cu.o
Scanning dependencies of target sift_gpu
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/FrameBufferObject.cpp.o
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/GlobalUtil.cpp.o
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/GLTexImage.cpp.o
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/ProgramGLSL.cpp.o
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/PyramidGL.cpp.o
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/ShaderMan.cpp.o
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/SiftGPU.cpp.o
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/SiftMatch.cpp.o
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/SiftPyramid.cpp.o
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/CuTexImage.cpp.o
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/PyramidCU.cpp.o
[ %] Building CXX object lib/SiftGPU/CMakeFiles/sift_gpu.dir/SiftMatchCU.cpp.o
[ %] Linking CXX static library libsift_gpu.a
[ %] Built target sift_gpu
Scanning dependencies of target sqlite3_automoc
[ %] Automatic rcc for target sqlite3
[ %] Built target sqlite3_automoc
Scanning dependencies of target sqlite3
[ %] Building C object lib/SQLite/CMakeFiles/sqlite3.dir/sqlite3.c.o
[ %] Linking C static library libsqlite3.a
[ %] Built target sqlite3
Scanning dependencies of target vlfeat_automoc
[ %] Automatic rcc for target vlfeat
[ %] Built target vlfeat_automoc
...

colmap_build

参考资料

[1] NVIDIA DEVELOPER

[2] CUDA TOOLKIT DOCUMENTATION

[3] Linux(Ubuntu)系统查看显卡型号

[4] 最全面解析 Ubuntu 16.04 安装nvidia驱动以及各种错误

[5] Ubuntu安装和卸载CUDA和CUDNN

[6] Ubuntu server16.04安装配置驱动418.87、cuda10.1、cudnn7.6.4.38、anaconda、pytorch超详细解决

[7] Ubuntu 16.04 安装 CUDA10.1 （解决循环登陆的问题）

[8] 【目标检测】Ubuntu16.04+RTX2070+CUDA10.0+pytorch1.1搭建CenterNet环境

Ubuntu16.04+GTX2070+Driver418.43+CUDA10.1+cuDNN7.6的更多相关文章

深度学习环境配置:Ubuntu16.04下安装GTX1080Ti+CUDA9.0+cuDNN7.0完整安装教程（多链接多参考文章）
本来就对Linux不熟悉,经过几天惨痛的教训,参考了不知道多少篇文章,终于把环境装好了,每篇文章或多或少都有一些用,但没有一篇完整的能解决我安装过程碰到的问题,所以决定还是自己写一篇我安装过程的教程, ...
Ubuntu16.04下nvidia驱动+nvidia-docker+cuda9+cudnn7安装
一.宿主机安装nvidia驱动打开终端,先删除旧的驱动: sudo apt-get purge nvidia* 禁用自带的 nouveau nvidia驱动 sudo gedit /etc/modp ...
（解决某些疑难杂症）Ubuntu16.04 + NVIDIA显卡驱动 + cuda10 + cudnn 安装教程
一.NVIDIA显卡驱动打开终端,输入: sudo nautilus 在新打开的文件夹中,进入以下路径(不要用命令行): 左下角点计算机,lib,modules 这时会有几个文件夹,对每个文件夹都进 ...
ubuntu16.04 安装NVIDIA和CUDA9.2 cudNN7.1
1.安装NVIDIA驱动 (1)查询NVIDIA驱动首先去官网(http://www.nvidia.com/Download/index.aspx?lang=en-us)查看适合自己显卡的驱动(下载 ...
【MindSpore】Ubuntu16.04上成功安装GPU版MindSpore1.0.1
本文是在宿主机Ubuntu16.04上拉取cuda10.1-cudnn7-ubuntu18.04的镜像,在容器中通过Miniconda3创建python3.7.5的环境并成功安装mindspore_g ...
ubuntu16.04配置tensorflow-gpu环境
1.安装驱动参考: 史上最全的ubuntu16.04安装nvidia驱动+cuda9.0+cuDnn7.0 https://blog.csdn.net/qq_31215157/article/det ...
Ubuntu16.04 + CUDA9.0 + cuDNN7.3 + Tensorflow-gpu-1.12 + Jupyter Notebook 深度学习环境配置
目录一.Ubuntu16.04 LTS系统的安装二.设置软件源的国内镜像 1. 设置方法 2.关于ubuntu镜像的小知识三.Nvidia显卡驱动的安装 1. 首先查看显卡型号和推荐的显卡驱动 ...
Ubuntu16.04安装cuda9.0+cudnn7.0
Ubuntu16.04安装cuda9.0+cudnn7.0 这篇记录拖了好久,估计是去年6月份就已经安装过几遍,然后一方面因为俺比较懒,一方面后面没有经常在自己电脑上跑算法,比较少装cuda和cudn ...
ubuntu16.04+cuda9+cudnn7+tensorflow+pycharm环境搭建
安装环境:ubuntu16.04+cuda9+cudnn7+tensorflow+pycharm 1)前期搭建过程主要是按照这篇博文,对于版本选择,安装步骤都讲得很详细,亲测有效! https://b ...

随机推荐

UDP 协议的那点事儿
最近在回顾计算机网络的知识,以前上课没有认真学,只记得几个高大上的术语,所以趁着这次回顾,把学到的知识用博客的形式记录下来,一来加深自己的印象,二来希望让你对这些基础知识有一个更深入的了解.当然,我会 ...
API规范约定
为了高效开发,节约编写文档的成本,API服务使用Swagger来描述一.API设计原则控制API的粒度和数量命名要遵循简单.可读.统一原则: 优先设计API,然后编码二.URL设计[针对后端开 ...
usermod命令、用户密码管理、mkpasswd命令使用介绍
第3周第2次课(4月3日) 课程内容:3.4 usermod命令3.5 用户密码管理3.6 mkpasswd命令 3.4 usermod命令 usermod可以修改用户的UID和GID 命令使用格式: ...
vim介绍、颜色显示和移动光标、一般模式下移动光标及复制、剪切和粘贴
第4周第4次课(4月12日) 课程内容: 5.1 vim介绍5.2 vim颜色显示和移动光标5.3 vim一般模式下移动光标5.4 vim一般模式下复制.剪切和粘贴 5.1 vim介绍 centos7 ...
split分割（拆分）文件
split分割(拆分)文件需求:指定文件大小拆分文件 # ll -h test/ |grep vmcore -rw-r--r-- 1 root root 12G 12月 7 00:20 vmco ...
MyBatis开发Dao的原始Dao开发和Mapper动态代理开发
目录咳咳...初学者看文字(Mapper接口开发四个规范)属实有点费劲,博主我就废了点劲做了如下图,方便理解: 原始Dao开发方式 1. 编写映射文件 3.编写Dao实现类 4.编写Dao测试 Ma ...
redis数据类型--hash
/** Redis应用之Hash数据类型* 问题1:操作命令* 问题2:存储实现原理和数据结构* 问题3:应用场景* */ 先了解下什么是hash,什么是hash碰撞:hash:是包含键值对的kv的数 ...
docker的安装及常用命令
一:概述 Docker 是一个开源的应用容器引擎,让开发者可以打包他们的应用以及依赖包到一个可移植的容器中,然后发布到任何流行的Linux机器或Windows 机器上,也可以实现虚拟化,容器是完全使用 ...
【原创】（十三）Linux内存管理之vma/malloc/mmap
背景 Read the fucking source code! --By 鲁迅 A picture is worth a thousand words. --By 高尔基说明: Kernel版本: ...
[TimLinux] JavaScript AJAX接收到的数据转换为JSON格式
1. 接收数据 AJAX接收数据是通过xhr.responseText属性,这是一个属性不是一个方法,这个属性得到的数据为字符串. 2. 字符串内容当服务器发送的是一个JsonResponse({' ...

Ubuntu16.04+GTX2070+Driver418.43+CUDA10.1+cuDNN7.6

Ubuntu16.04+GTX2070+Driver418.43+CUDA10.1+cuDNN7.6的更多相关文章

随机推荐

热门专题