Skip to content

user_guide.md

Latest commit

 

History

History
218 lines (153 loc) · 6.25 KB

user_guide.md

File metadata and controls

218 lines (153 loc) · 6.25 KB

Easley

Initial setup and connection

  1. Request a user account. You should select Howard as your "sponsor". You will get an email confirmation after you account is approved by me and the HPC admins.

  2. Connect to the VPN if you are off campus or using WiFi (regardless of location). You will want to make sure that your Duo 2-factor authentication (2FA) is setup to use push notifications.

  3. Connect to Easley.

    If you are a Unix (Mac or Linux) user, open a terminal and connect with ssh:

    ssh <auburn-username>@easley.auburn.edu

    You will be prompted to enter your login credentials (including 2FA), then you will arrive at a splash screen in your home directory. To simplify the connection process in future, you can add Easley as a Host entry in your ~/.ssh/config file:

    Host easley
    HostName easley.auburn.edu
    User <auburn-username>

    so that you can use the shorter command:

    ssh easley

    If you are a Windows user, download PuTTY. Enter easley.auburn.edu as the hostname, and connect to port 22 with ssh. You will be asked for your login credentials, then you will get a terminal similar to the Unix users.

  4. In order to use the group software and handle some common configuration tasks, run the following from your command prompt:

    ./home/shared/mph0043_lab/software/config_bash.sh

    This command should only be run once (the first time you login), and does not need to be repeated in future.

  5. Log out using the logout command. The changes you have made will take effect the next time you login.

Software

Easley has a large set of software already installed for you, typically with multiple versions. The software that is available may change depending on what else you have already loaded (for example, the compiler you are using). The packages are managed using the module tool. To see what software has already been loaded, use:

module list

By default, a few things will already be loaded for you when you log in. To see all the software that is available:

module avail

You can also check for specific versions of a particular package:

module avail python

To load a particular software and make it available for use:

module load python/3.9.2

If you run python --version, you should see Python 3.9.2.

To deactivate it:

module unload python/3.9.2

and running python --version will give you the system python (currently Python 2.7.5).

If a package is marked with a (D), it is the "default" that will be loaded if you don't specify the full package name. If a version is marked with an (L), it is currently loaded. For convenience, you can tab complete names of modules so that you don't have to type the full name manually.

Our group also has software installed for everyone to use. You should see it at the top when you do module list, under the heading /home/shared/mph0043_lab/software/modulefiles. You can activate this software in the same way as any of the other modules.

If you get yourself in a mess with your modules, you can always log out and log back in again to get a clean start!

Using Anaconda Python

Initial setup

  1. Load a specific version of Anaconda. This example uses 3.8.6:

    module load python/anaconda/3.8.6

  2. To make sure the shared Anaconda installation can create environments correctly, you need to create a ~/.condarc file and add the following:

    pkgs_dirs:
 - ~/.conda/pkgs
 - /tools/anacondapython-3.8.6/pkgs

    The white space on lines 2 & 3 is important, so make sure you preserve it: [space][-][space][rest of the line]. In line 3, the version number should match the one you loaded. After creating this file, run the command conda info and you should see two entries in the "package cache" entry.

Creating environments

You can now create new conda environments that get installed in ~/.conda/envs:

conda create -n testenv

You can also create new environments at a specific location:

conda create -p /path/to/env

Note that there are many additional options for conda create. For example, you can set the version of Python to use in the environment. Refer to the Anaconda documentation for more details.

Using environments

You can activate / deactivate these environments using source activate:

source activate testenv

or

source activate /path/to/env

You should now see the name of the environment preprended to your command prompt like (testenv).

You may get a suggestion to run conda init instead so that you can activate the environment using conda activate. However, this command modifies your .bashrc file, so I don't recommend it when source activate works just fine.

Installing in an environment

After you activate an environment, you can use conda or pip to install new software. This gives you a nice way to create isolated environments with compatible versions of software. You should always prefer to install from conda before pip when using Anaconda in order to get the best dependency management.

Use conda install to install new packages. You may need to add a channel like conda-forge:

conda install -c conda-forge hoomd

Note that when using conda-forge, it can take a long time to solve for an environment. You may need to be patient in these cases, or reconsider what you are installing into the the same environment.

To uninstall a package, use conda remove:

conda remove hoomd

Deactivating and removing an environment

Regardless of how the environment was activated, you deactivate it the same way:

conda deactivate

To remove an environment created by name (-n), first deactivate it then use:

conda remove -n testenv --all

To remove an environment created by prefix, first deactivate it then delete it:

rm -r /path/to/env