Linux Essentials

Table of contents:

• Linux Essentials
  • General
  • Help
  • Audio/Video
  • Working with Files
  • Compression
  • Search for Information / Information Extraction from Files
  • REGEX
  • Pipe
  • Bash
  • Main Directories in Linux
  • Processes
  • Networking
    • How to configure static ip address
  • Users
  • Permissions
  • SYMBOLIC LINKS (EQUIVALENT OF SHORTCUTS IN WINDOWS)
  • Sticky Bit
  • AWS
  • Anaconda
  • GPU
    • How to Install CUDA
  • Screen (Do Something while I'm gone)
  • Server
  • Common Problems

General

Back to top

command	what it does
clear	clear the terminal
history	list all of the commands we executed
du -hs /path/to/a/directory/	see the volume of a directory
df -h	see the disk space left for all of the disks
nvtop	see the status of the NVIDIA GPU
htop	see the status of the CPUs and RAM
pwd	show the current directory that the commands are being executed in
rm -rf ~/.local/share/Trash/*	Empty the trash
lsof / | grep deleted	List the files that should have been deleted, but are not deleted until now!
lsof|grep deleted|awk '{print $2}'|xargs kill -9	Get rid of the deleted files for which a job/process is still open, preventing the file from being deleted
pip cache purge	clear the cache for pip

Help

Back to top

ls --help	documentation of the "ls" command
man echo	manual of the "echo" command
info echo	more information about the "echo" command (Note: More complete than "man" and "help" commands)
whatis ls	what is "ls"? (a very short answer to this)

Audio/Video

Back to top

ffmpeg -i whatever.webm -c copy -strict experimental whatever.mp4	convert webm file to mp4 (Note: "-strict experimental" was added due to the outdated ffmpeg package)
ffmpeg -fflags -genpts -i whatever.webm -r 24 whatever.mp4	convert webm file to mp4 (Another method)
ffmpeg -i input.mp3 -acodec pcm_s16le -ac 1 -ar 16000 out.wav	convert 'input.mp3' to a wav file named 'out.wav' (16000 and mono)

Working with Files

Back to top

sudo du -aBm / 2>/dev/null | sort -nr | head -n 10	list the biggest files/directories in terms of volume (top 10)
du -hsx * | sort -rh | head -10	list the biggest files/directories in terms of volume (top 10)
find . -type f -printf '%s %p\n'| sort -nr | head -10 | numfmt --field=1 --to=iec	list the largest files in terms of volume (top 10)
cd /the/path/of/a/folder/	change directory
cd ~	change directory to the home dir
cd /	change directory to the root dir
ls	list the files in the current directory
ls -a	list the files in the current directory (including the hidden files) (-a means all) - the files having . at the beginning of their names are hidden
ls -l	list the files in the directory with more information (including their read/write access, owner, volume, date&time of creation)
ls -la	list ALL of the files in the directory with more information (including their read/write access, owner, volume, date&time of creation)
ls /path/to/some/directory/	list the files in a given directory
ls mypattern*	list the files in the current directory with names starting with "mypattern"
ls *mypattern	list the files in the current directory with names ending with "mypattern"
rm -rf /directory/	remove everything inside a directory ("r" is for recursive deletion, "f" removes read-only files without asking
mkdir directory_name	make a new directory (folder)
touch test.txt	make a new file named "test.txt" in the current directory
nano test.txt	edit the "test.txt" file in terminal
cat test.txt	print the content of "test.txt" in the terminal
rm test.txt	remove the file "test.txt"
rm porteqal.mp4 sound.mp3 linux.png	remove the three files "porteqal.mp4", "sound.mp3", and "linux.png"
rmdir /mydirectory/	remove the directory "mydirectory" (NOTE: works if it's empty)
sudo chmod +x myscript.sh	make the script "myscript.sh" executable
x=2	define the variable "x" to be 2
echo $x	print the content of the variable "x"
echo "salam"	print the word "salam"
echo $a $b	print the content of the variable "a", then a space, and then the content of the variable "b"
scp -r speaker* [email protected]:~/wavfiles	Copy the folders whose name start with "speaker" from here to the server 120.233.50.2, while logging into that server with the username "arman"
scp -r /Users/armanmalekzadeh/Documents/BERT-NEW/ [email protected]:~/arman/BERT-Refactor/codes/training/bert-checkpoints/	Copy the folders and files inside the "BERT-NEW" directory on my own laptop to the "bert-checkpoints" directory on a remote server
cp test.txt /path/to/a/directory	copy the file "test.txt" to a directory
cp -r * /path/to/a/directory/	copy everything inside the current directory (including the folders) to another directory
cp -r /first/dir/ /second/dir/	copy the first directory into the second directory
cp -a /source/. /dest/	copy everything from the "source" to the "dest" folder
rsync -vau --remove-source-files source/ dst/	Move all files from the "source/" directory to the "dst/" directory (Note: It's sometimes faster than the "cp"/"mv" command)
mv test.txt /path/to/a/directory/	move the file "test.txt" to another directory
mv *.txt /path/to/a/directory/	move all of the files whose name end with ".txt" to another directory
mv a.txt b.txt	rename "a.txt" to "b.txt"
whereis ls	show the directory in which the "ls" library exists
less a.txt	read the contents of a text file one page(one screen) at a time.

Compression

Back to top

zip media.zip porteqal.mp4 sound.mp3 linux.png	make a zip file containing the tree files "porteqal.mp4", "sound.mp3", and "linux.png". Name the file "media.zip"
unzip media.zip	decompress the file "media.zip", and move its contents to the current directory
zip linux.zip /path/to/a/directory/	make a compressed file containing a directory itself (without the files it contains!!!)
zip -r tutorial.zip /path/to/a/directory/	make a compressed file containing a directory and its contents
unzip media.zip -d /path/to/a/directory/	decompress the file "media.zip", and move its contents to a desired directory
tar -cf myfile.tar porteqal.mp4 sound.mp3 linux.png	"create" a tar "file" (a compressed file with no volume decreasion), given a list of files
tar -xf myfile.tar	decompress (extract) a tar file
tar -vxf myfile.tar	decompress (extract) a tar file while printing out the things inside it
gzip myfile.tar	compress the given "tar" file (which is not volume-efficient) to a .tar.gz compressed file (more compressed than the tar file)
gunzip myfile.tar.gz	decompress the .gz file into the .tar file
bzip2 myfile.tar	compress the given "tar" file into a .tar.bz2 compressed file (more compressed than "tar")
bunzip2 myfile.tar.bz2	decompress the .bz2 file into the .tar file
tar -cvzf media.tgz sound.mp3 porteqal.mp4	compress some files directly to a tar gzip file (instead of first compressing as "tar", and then "gzip") - Note: "v" stands for "verbose mode"
tar -cvjf media.bz2 sound.mp3 porteqal.mp4	compress some files directly to a tar bzip file (instead of first compressing as "tar", and then "bzip")
tar -xvzf media.tgz	Decompress the .tgz file directly into the original files
tar -xvjf media.bz2	Decompress the .bz2 file directly into the original files

Search for Information / Information Extraction from Files

Back to top

less myfile.txt	displays the content of a file "page by page" with scrolling option
head myfile.txt	Displays the first 10 lines of a file
tail myfile.txt	Displays the last 10 lines of a file
head -n 15 myfile.txt	Displays the first 15 lines of a file
tail -n 15 myfile.txt	Displays the last 15 lines of a file
sudo find /	Search in the whole files and directories recognized by the OS for files, and list them
find Downloads/ -type f	Search for files (not folders) in the "Downloads" folder (and list them!)
sudo find / -name "python"	Search for files and directories named "python", and list them
sudo find / -name "python*"	Search for files and directories whose name starts with "python"
grep arman myfile.txt	Search for the expression "arman" inside the file "myfile.txt", and if it exists, print the token containing it (e.g., print "arman.ir") - Note: If the expression doesn't exist, the output of this command will be empty!
grep arman *	Search for the expression "arman" inside all of the files inside the current directory, and show the results (as a list)
grep "arman malekzadeh" *	Search for the long expression "arman malekzadeh" inside all of the files inside the current directory, and show the results (as a list)
sort myfile.txt	Display the content of the file "myfile.txt" (with lines sorted alphabetically!) - Note: Each line is considered as a separate string
sort -n myfile.txt	Display the content of the file "myfile.txt" (with lines sorted based on numbers) - Note: Each line is considered as a separate string
sort -r myfile.txt	Display the content of the file "myfile.txt" sorted in the reverse order
sort -R myfile.txt	Display the content of the file "myfile.txt" sorted randomly!
cut -c3-6 myfile.txt	Display the content of the file "myfile.txt", but only columns 3 to 6 for each line
cut -c3,6 myfile.txt	Display the content of the file "myfile.txt", but only the third and sixth character for each line
cut -c3- myfile.txt	Display the content of the file "myfile.txt", but only the third character and everything after it, for each line
cut -d" " -f4 myfile.txt	Split the contents of a file using the space character, consider each part as a "field", and display the content of the fourth field - Note: this is useful for reading "csv" files
cut -d" " -f4- myfile.txt	Split the contents of a file using the space character, consider each part as a "field", and display the content of the fourth field, and every field after it
cut -d" " -f4-6 myfile.txt	Split the contents of a file using the space character, consider each part as a "field", and display the content of the fourth field up to the sixth field
cut -d" " -f4,6 myfile.txt	Split the contents of a file using the space character, consider each part as a "field", and display the content of the fourth field and the sixth field
cut -d" " -f3-6 myfile.txt > output.txt	Split the contents of a file using the space character, consider each part as a "field", and display the content of the third field up to the sixth field. Finally, write the output into the file "output.txt"
cut -d" " -f3-6 myfile.txt >> output.txt	Split the contents of a file using the space character, consider each part as a "field", and display the content of the third field up to the sixth field. Finally, append the output to the content of the file "output.txt"
wc myfile.txt	Display the number of lines that the file contains, then the number of words (tokens separated by space), and the number of characters

REGEX

Back to top

grep ^arman myfile.txt	Search for the lines starting with "arman" inside "myfile.txt", and show the results (as a list)
grep arman$ myfile.txt	Search for the lines ending with "arman" inside "myfile.txt", and show the results (as a list)
grep a.m myfile.txt	Search for the lines starting with "a", having "m" as their third character, (don't care about the second character) inside "myfile.txt", and show the results (as a list)
grep a..m myfile.txt	Search for the lines starting with "a", having "m" as their fourth character, (don't care about the second and third characters) inside "myfile.txt", and show the results (as a list)
grep ^....$ myfile.txt	Search for the lines having only 4 characters inside "myfile.txt", and show the results (as a list)

Pipe

Back to top

command1 | command2	run the first command, and take its output to the second command as its input
ls | grep ^D	/ list the files of the current directory, but only show the results starting with the "D" character
ls | grep s$	list the files of the current directory, but only show the results ending with the "s" character
find . type f | grep python	find the files in the system named "python"
cut -d" " -f3 myfile.txt | cut -c2	display the second character of the third token from each row

Bash

Back to top

ls sh	get the path to the interpreter of the bash language
#!/bin/bash your bash code goes here	how to write a shell script with bash
#! /etc/bash echo "please write your name" if [[ "$1" == "arman" ]]; then     echo "yes" else      echo "no" fi	a demo bash script using the conditional statement "if"
sudo chmod 777 myscript.sh	Give the full permission to everyone to do everything with the file "myscript.sh" (including the permission to run it)
bash ./myscript.sh	run my shell script named "myscript.sh"
bash ./myscript.sh arman	run my shell script named "myscript.sh" taking "arman" as input
#! /etc/bash for i in {1..10} do     echo $i done	print the numbers 1 to 10
#! /etc/bash for (( i=1 ; i <= 50 ; i++ )) do      echo "number : $i$" done echo "done"	write the numbers from 1 to 50 with a "number : " string on the left side of them

Main Directories in Linux

Back to top

etc	all of the linux tools are here, including bluetooth, apache, python, and everything you can install on it. So, if you wanna configure sth, here is the place to look for the .conf file
bin	contains the code for all of the commands such as "ls"
lib	contains the libraries
/var/log	the logs in linux are here
dmsg	show the logs in a beautiful way (for instance, if you connect a usb, this will show sth)

Processes

Back to top

ps	list the processes that are running
ps -aux	show the list of the processes being run by any user, along with the process IDs (PIDs) (Note: aux stands for all user execute)
kill 5577	kill the process with ID 5577
top	show the processes that are being run (live!)
free	show the ram (memory), how much of it is used, how much is free, and the same things for the swap (which is used to accelerate the speed of the system)
free -tb	show the ram (memory), how much (how many bytes) of it is used, how much is free, and the same things for the swap (which is used to accelerate the speed of the system)
free -tm	show the ram (memory), how much (how many megabytes) of it is used, how much is free, and the same things for the swap (which is used to accelerate the speed of the system)

Networking

Back to top

/etc/resolv.conf	the file containing the configure of the DNS It may contain sth like this: nameserver 127.0.0.25 options edns0 search localdomain To add another DNS (like the DNS server of the google which is 8.8.8.8), append the following line: nameserver 8.8.8.8
ifconfig	shows some information about the conifiguration of the network (such as the MAC address), the networking cards, ... (Note: "if" stands for "interface")
ifconfig -a	show the network cards including the disabled ones
sudo ifconfig eth0 192.168.245.120	change the ip address of the network card "eth0" to 192.168.245.120
sudo ifconfig eth0 down	disable the network card "eth0"
sudo ifconfig eth0 up	enable the network card "eth0"
dig arman.ir	reveal some information about the website "arman.ir" including the IP addresses it has - Actually, this command sends a request to the nameservers of the website
ping 192.168.200.567	send requests to the given ip address to see if we can have a connection with the device (pc) having that ip address - Note: in the result of this command, you will see "TTL". If the value of "ttl" is more than 100, the pc has a windows installed on it. Otherwise, the OS is linux. - Note2: You are not required to write an ip address. Instead, you can write a domain, such as ping arman.ir - Note3: This command is used to see if the connection between two clients or a client and a server has been established

How to configure static ip address

1. Identify the name of the ethernet interface using ip link (or ip a, ip addr). The output will be something like this:

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: eno1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
    link/ether 0c:c4:7a:cc:59:70 brd ff:ff:ff:ff:ff:ff
3: eno2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
    link/ether 0c:c4:7a:cc:59:71 brd ff:ff:ff:ff:ff:ff

Here the name is eno1. It can be ens1, eth0 or sth like that in other cases.

2. Go to the /etc/netplan/ directory. There are some files with the .yaml postfix. Edit those files and make them like this:

network:
  version: 2
  renderer: networkd
  ethernets:
    eth0:
      dhcp4: no
      addresses:
        - 192.168.121.199/24
      gateway4: 192.168.121.1
      nameservers:
          addresses: [8.8.8.8, 1.1.1.1]

In the above lines, eth0 is the name of your interface, 192.168.121.199/24‍‍‍ is the static ip address. If you want multiple addresses, type another line for the other address.

3. Test the connection to the network (if you want to) using sudo netplan try.

4. Apply the new network plan by running sudo netplan apply.

5. Verify the changes by running ip addr show dev eno1. Remember that eno1 was the name of our interface. The output will be something like this:

3: eno1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
    link/ether 56:00:00:60:20:0a brd ff:ff:ff:ff:ff:ff
    inet 192.168.121.199/24 brd 192.168.121.255 scope global dynamic eno1
       valid_lft 3575sec preferred_lft 3575sec
    inet6 fe80::5054:ff:feb0:f500/64 scope link 
       valid_lft forever preferred_lft forever

---

Note: For older Ubuntu versions (e.g., 16), there might be no /etc/netplan/ directory. In this case, check for the /etc/network/interfaces file. It should look like this:

# interfaces(5) file used by ifup(8) and ifdown(8)
auto eno1
iface eno1 inet static
 address 213.220.120.2 (ip address of the server)
 netmask 255.255.255.0 (always the same)
 gateway 213.220.180.1 (ask this from the server admin)
 dns-nameservers 4.4.4.4 8.8.8.8 (always the same)

After doing this, restart the network service by running sudo /etc/init.d/networking restart.

Users

Back to top

who	shows the current user, with a little bit of extra info
w	shows which users are logged in, the time of their login, and what they are doing
whoami	shows the current user's name
id	shows the unique id of the current user, the groups he/she belongs to - Note: From 0 to 499, is for system users, and 500 to 6000 is for standard users
id anotheruser	shows the unique id of the given user named "anotheruser", the groups he/she belongs to
cat /etc/passwd	the file /etc/passwd contains information about all of the users, including the uid of them and the id the group they belong to, and the home directory (a.k.a profile) (for the standard users)
cat /etc/group	the file /etc/group contains information about all of the groups, including the id of each group and the users belonging to each group
sudo cat /etc/shadow	the file /etc/shadow contains the hashed passwords for all users
sudo command	execute the command using the permission level of the root user
su arman	change the user to arman in order to execute commands on his/her behalf (this will ask for the password of the user arman)
sudo adduser arman	create a new user named arman - Note: this will ask for a password for the new user, and some other info (such as phone number!)
sudo addgroup mynewgroup	create a new group named mynewgroup
sudo passwd	change the password of the current user
sudo passwd arman	change the password of the user named arman
last	shows the last logins in the system
sudo usermod -a -G mygroup arman	Add the user arman to the group mygroup ("a" stands for "add", "G" for "group")
sudo usermod -a -G root arman	Add the user arman to the root group
sudo userdel arman	delete the user arman, but keep his/her profile - Note: the path /home/arman/ will remain.
sudo userdel -r arman	delete the user arman and his/her profile (files)
sudo groupdel mygroup	delete a group named mygroup
sudo adduser arman sudo	add user arman to the sudo group
usermod -aG sudo arman	add user arman to the sudo group
sudo deluser arman sudo	revoke sudo access from arman
grep -Po '^sudo.+:\K.*$' /etc/group	get the list of sudoers

Permissions

Back to top

ls -l	The first character: "d" means directory, and "-" means file The three characters after that: "r" = read, "w" = write, "x" = execute. If instead of any of them, a "-" is written, means that we cannot "read", "write" or "execute" that file. The first three characters correspond to the permissions of the owner of the file The second three characters correspond to the group members of the owner The last three characters correspond to other users (any kind of user) The "execute" permission for directories means changing directory to those dirs (using the cd command)
About the numbers of the characters mentioned above	"r" (read) = 4, "w" (write) = 2, "x" (execute) = 1 - Note that the sum of them is 7. So, 777 means that anyone can do anything with that file/directory
sudo chmod u	give the permission of ... to the user named ``
sudo chmod o	give the permission of ... to the other users named ``
sudo chmod g+w myfile.txt	give the permission to write ("w") to the file "myfile.txt" to the group members of the owner of the file
sudo chmod uog+w myfile.txt	give the permission to write ("w") to the file "myfile.txt" to everyone
sudo chmod g+rw myfile.txt	give the permission to read and write ("rw") to the file "myfile.txt" to the group members of the owner of the file
sudo chmod g+rxw myfile.txt	give the permission to read, execute and write ("rxw") to the file "myfile.txt" to the group members of the owner of the file
sudo chmod g+rxw myfile.txt	give the permission to read, execute and write ("rxw") to the file "myfile.txt" to the group members of the owner of the file
sudo chmod gou-wx myfile.txt	take away the permissions to write to and execute the file "myfile.txt" from every user
sudo chmod ug+w myfile.txt	give the permission to write to the file "myfile.txt" to the owner and the group it belongs to
sudo chmod 700 myfile.txt	give the permission to do anything to the owner of the file "myfile.txt", and do nothing to others
sudo chmod +x myfile.txt	give the permission to write to the file "myfile.txt" to everyone!
sudo chown arman myfile.txt	change the owner of the file "myfile.txt" to the user "arman"
sudo chgrp mygroup myfile.txt	change the group of the file "myfile.txt" to the group "mygroup"
sudo chown arman myfolder -R	change the owner of the directory named "myfolder" to the user "arman", along with all of its subfolders and files inside it
sudo chmod -R 777 /path/to/your/directory	give every user the permission to do everything with the /path/to/your/directory and all its contents (including files and folders)

SYMBOLIC LINKS (EQUIVALENT OF SHORTCUTS IN WINDOWS)

Back to top

ln -s /path/to/the/original/file.txt myshortcut.txt	create a shortcut of the file which is located at /path/to/the/original/file.txt in the current directory, and name it as myshortcut.txt

Sticky Bit

Back to top

sudo chmod o+t myfolder/	every user can only change the files that he/she owns
sudo chmod 777 myfolder	remove the sticky bit mode

AWS

Back to top

aws s3 cp /path/to/a/file.txt s3://[bucket-name]/path/to/file.txt --endpoint-url https://[bucket-name].parspack.net	copy a file to a bucket
aws s3 cp /path/to/a/folder s3://[bucket-name]/ --recursive --endpoint-url https://[bucket-name].parspack.net	copy a folder to a bucket
aws s3 ls s3://[bucket-name]/path/to/a/folder/ -endpoint-url https://[bucket-name].parspack.net	list the contents of a folder

Anaconda

Back to top

source ~/anaconda3/bin/activate	activate the Anaconda's environment
conda create --name your_desired_name python=3.8	make a new environment which uses a specific Python version
conda env list	list all environments built inside anaconda
conda env remove --name ENVIRONMENT_NAME	remove the conda environment named "ENVIRONMENT_NAME"
jupyter notebook --no-browser --port=8080	Run a jupyter notebook on the server without using the browser on the port 8080
jupyter notebook password	Define a new password for the jupyter notebook (instead of the token, you will give this password to anyone who will use the jupyter)
jupyter notebook --no-browser --port=8080 --notebook-dir=/home/user01/	Run a jupyter notebook on the server (at path /home/user01/ without using the browser on the port 8080
tensorboard --logdir /path/to/save/the/logs --port 8895	Run Tensorboard on port 8895 and save its logs to the specified directory (logdir)
ssh -L 8080:localhost: <REMOTE_USER>@<REMOTE_HOST>	Use a jupyter which is running on port <PORT> on your local port 8080 (by connecting to the <REMOTE_USER>@<REMOTE_HOST> using SSH)

Note: To install Anaconda for all linux users, do the following steps:

1. Download Anaconda .sh installation file from its website.
2. Run sudo -i to be the root user.
3. Run sudo apt-get update
4. Execute bash Anaconda[...].sh to run the installation file.
5. Accept the license (type yes)
6. For the install location, choose /opt/anaconda3
7. It asks "Do you wish the installer to initialize blah blah blah"? Say "yes". (The installation is finished).
8. Run exit to again be a normal user.
9. Now your user does not recognize conda (For instance if you type conda -V, it will say command not found).
10. Run sudo nano /etc/environment.
11. At the beginning of the PATH variable, after the " character, add /opt/anaconda3/bin:, and save the file using CTRL+O.
12. Log out from your own user and do the SSH again. Now if you run conda -V, the user will recognize it.

GPU

Back to top

Installing CUDA 11.2 and cuDNN 8.1	Unofficial Link - NVIDIA Link for cuDNN Installation - How to verify cuDNN Installation

How to Install CUDA

• sudo apt-get update
• sudo apt-get install -y build-essential cmake unzip pkg-config
• sudo apt-get install -y libxmu-dev libxi-dev libglu1-mesa libglu1-mesa-dev
• sudo apt-get install -y libjpeg-dev libpng-dev libtiff-dev
• sudo apt-get install -y libavcodec-dev libavformat-dev libswscale-dev libv4l-dev
• sudo apt-get install -y libxvidcore-dev libx264-dev
• sudo apt-get install -y libgtk-3-dev
• sudo apt-get install -y libopenblas-dev libatlas-base-dev liblapack-dev gfortran
• sudo apt-get install -y libhdf5-serial-dev graphviz
• sudo apt-get install -y python3-dev python3-tk python-imaging-tk
• sudo apt-get install -y linux-image-generic linux-image-extra-virtual
• sudo apt-get install -y linux-source linux-headers-generic
• sudo apt-get purge nvidia* (remove any nvidia drivers already installed)
• sudo add-apt-repository ppa:graphics-drivers/ppa (add graphic drivers ppa)
• sudo apt-get update
• ubuntu-drivers devices (search for available drivers). The output will be something like this:

== /sys/devices/pci0000:00/0000:00:03.0/0000:03:00.0 ==

vendor   : NVIDIA Corporation

model    : GM200 [GeForce GTX TITAN X]

modalias : pci:v000010DEd000017C2sv000010DEsd00001132bc03sc00i00

driver   : nvidia-415 - third-party free

driver   : nvidia-430 - third-party free recommended

driver   : nvidia-418 - third-party free

driver   : nvidia-410 - third-party free

driver   : nvidia-384 - distro non-free

driver   : xserver-xorg-video-nouveau - distro free builtin

• sudo apt-get install nvidia-driver-430 (install the driver with the best version). The result was unable to locate package. Run sudo apt-get install nvidia-430 instead, and it will work!
• sudo reboot (before the reboot, nvidia-smi won't work. It will say something other than command not found which means a driver is there, but it will not work)
• To install tensorflow for a conda env: First activate it, and then run conda install -c anaconda tensorflow-gpu (install tensorflow-gpu) (while installing this, cuda-toolkit and cudnn will also be installed)
• apt install nvidia-cuda-toolkit (run as root and the base conda env)
• Run nvcc -V to get the version of the cuDNN! (Now it should work!)

Screen (Do Something while I'm gone)

Back to top

screen -S arman	Start a screen with name arman. You can detach from the screen (get out of it) by using ctrl+a and then pressing d
screen -r arman	Resume a screen named arman
screen killall	Kill all screens
screen -ls	List all running screens
screen -r -d 29284	Resume a screen with ID 29284

Server

Back to top

apt install ubuntu-desktop	install a desktop on an Ubuntu server

Common Problems

Back to top

sudo apt-get update fails

This is usually due to a corrupted sources.list.d file.

sudo mv /etc/apt/sources.list{,.bk}
sudo mkdir /etc/apt/sources.list.d
sudo cp /usr/share/doc/apt/examples/sources.list /etc/apt/sources.list
sudo apt-get update