- https://wiki.x2go.org/doku.php
- https://wiki.x2go.org/doku.php/doc:installation:x2goclient
The default installation of Miniconda may have restrictive licensing on some packages. The drop-in replacement "miniforge" will set up a similar minimal conda environment that defaults to the "conda-forge" channel of open packages.
As the packages in Python distributions can change rapidly, it is difficult to have a single, system-wide installation that is useful to everyone. We will continue to install some basic ones in /sopt, but they are not likely to change after initial build. To let users customize a minimal Python environment that won't disturb others, we suggest installing the "Miniconda" version of the Anaconda Python distro. (See [https://conda.io/miniconda.html](https://conda.io/miniconda.html) for additional info.) **Example Install 1)** Python3-based conda installed in the default $HOME/miniconda3/ directory: ```bash cd wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh chmod u+x Miniconda3-latest-Linux-x86_64.sh ./Miniconda3-latest-Linux-x86_64.sh -b ``` **Example Install A)** Same as option1, but installed to /DFS-L/DATA/$group/$user/miniconda3/ directory, with link in home directory (allows installing very large numbers of packages that would make your home directory go over quota): ```bash cd wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh chmod u+x Miniconda3-latest-Linux-x86_64.sh mkdir /DFS-L/DATA/$(id -gn)/$USER/miniconda3 ln -s /DFS-L/DATA/$(id -gn)/$USER/miniconda3 miniconda3 ./Miniconda3-latest-Linux-x86_64.sh -bu ``` As of 23-November-2020, this installs Miniconda3 4.9.2, which uses Python 3.8.5. It also installs the minimal set of packages: ```bash _libgcc_mutex-0.1-main brotlipy-0.7.0-py38h27cfd23_1003 ca-certificates-2020.10.14-0 certifi-2020.6.20-pyhd3eb1b0_3 cffi-1.14.3-py38h261ae71_2 chardet-3.0.4-py38h06a4308_1003 conda-4.9.2-py38h06a4308_0 conda-package-handling-1.7.2-py38h03888b9_0 cryptography-3.2.1-py38h3c74f83_1 idna-2.10-py_0 ld_impl_linux-64-2.33.1-h53a641e_7 libedit-3.1.20191231-h14c3975_1 libffi-3.3-he6710b0_2 libgcc-ng-9.1.0-hdf63c60_0 libstdcxx-ng-9.1.0-hdf63c60_0 ncurses-6.2-he6710b0_1 openssl-1.1.1h-h7b6447c_0 pip-20.2.4-py38h06a4308_0 pycosat-0.6.3-py38h7b6447c_1 pycparser-2.20-py_2 pyopenssl-19.1.0-pyhd3eb1b0_1 pysocks-1.7.1-py38h06a4308_0 python-3.8.5-h7579374_1 readline-8.0-h7b6447c_0 requests-2.24.0-py_0 ruamel_yaml-0.15.87-py38h7b6447c_1 setuptools-50.3.1-py38h06a4308_1 six-1.15.0-py38h06a4308_0 sqlite-3.33.0-h62c20be_0 tk-8.6.10-hbc83047_0 tqdm-4.51.0-pyhd3eb1b0_0 urllib3-1.25.11-py_0 wheel-0.35.1-pyhd3eb1b0_0 xz-5.2.5-h7b6447c_0 yaml-0.2.5-h7b6447c_0 zlib-1.2.11-h7b6447c_3 ``` This uses about 323MB of disk space. Once installed, you can set up the environment paths to your private version (with either install option) using: ```bash ml miniconda/3/own ``` Installing new packages (e.g. numpy) within your miniconda3 directory is as simple as: ```bash conda install numpy ``` Miniconda2 can be installed in a similar way with: ```bash cd wget https://repo.continuum.io/miniconda/Miniconda2-latest-Linux-x86_64.sh chmod u+x Miniconda2-latest-Linux-x86_64.sh ./Miniconda2-latest-Linux-x86_64.sh -b ``` There is also the corresponding miniconda/2/own module. Only one python module (miniconda/anaconda/python/Intel-python, all with versions 2 or 3) can be loaded at a time. # MiniforgeThe default installation of Miniconda may have restrictive licensing on some packages. The drop-in replacement "miniforge" will set up a similar minimal conda environment that defaults to the "conda-forge" channel of open packages.
As the packages in Python distributions can change rapidly, it is difficult to have a single, system-wide installation that is useful to everyone. We will continue to install some basic ones in /sopt, but they are not likely to change after initial build. To let users customize a minimal Python environment that won't disturb others, we suggest installing the "Miniforge" version of the Conda Python environment. (See [https://conda-forge.org/download/](https://conda-forge.org/download/) for additional info.) **Example Install 1)** Python3-based conda installed in the default $HOME/miniforge3/ directory: ```bash cd wget https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Linux-x86_64.sh chmod u+x Miniforge3-Linux-x86_64.sh ./Miniforge3-Linux-x86_64.sh -b ``` **Example Install A)** Same as option1, but installed to /X2/SCRATCH/$group/$user/miniforge3/ directory, with link in home directory (allows installing very large numbers of packages that would make your home directory go over quota): ```bash cd wget https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Linux-x86_64.sh chmod u+x Miniforge3-Linux-x86_64.sh mkdir /X2/SCRATCH/$(id -gn)/$USER/miniforge3 ln -s /X2/SCRATCH/$(id -gn)/$USER/miniforge3 miniforge3 ./Miniforge3-Linux-x86_64.sh -bu ``` As of 30-Marchr-2026, this installs Conda 26.1.1, which uses Python 3.13.12. This uses about 552MB of disk space. Once installed, you can set up the environment paths to your private version (with either install option) using: ```bash ml miniforge/3/own ``` Updating will upgrade a few packages ```bash conda update --all ``` Installing new packages (e.g. numpy) within your miniforge3 directory is as simple as: ```bash conda install numpy ``` You can create a new environment for CPU-based PyTorch: ```bash conda create --name torch-CPU pytorch torchvision torchaudio cpuonly -c pytorch ``` Or one for PyRosetta: ```bash conda create --name pyrosetta -c https://conda.rosettacommons.org ``` Only one python module (miniforge/miniconda/anaconda/python/Intel-python) can be loaded at a time. # JupyterHub While Jupyter notebooks can be run on any login node by installing the needed packages via conda or pip, running on compute nodes can be tricky. We have installed JupyterHub servers on gplogin2 and gplogin3 that can spawn JupyterLab sessions as slurm jobs. #### Starting Point your local computer's Web browser to either [https://gplogin2.ps.uci.edu:8000](https://gplogin2.ps.uci.edu:8000) or [https://gplogin3.ps.uci.edu:8000](https://gplogin3.ps.uci.edu:8000) and enter your Greenplanet username and password. Then, for the "simple" configuration: 1. Select which partition to run in 2. Select number of CPUs 3. Select which Jupyter environment to use (more info below) 4. If JupyterLab is desired, click box 5. Select runtime limit 6. Click Start For "Advanced", you have much more control of each option. See [https://github.com/silx-kit/jupyterhub\_moss](https://github.com/silx-kit/jupyterhub_moss) for detailed descriptions. #### Jupyter environments ##### Default The default Python environment (/sopt/JupyterHub/) is what is used to run the JupyterHub software. It is not user-customizable, but has a lot of general-use conda and pip packages and may just work. We can update and add more packages if needed at any time, so don't count on it having specific package versions. ##### Miniforge3 One of the pre-configured environments assumes you have miniforge3 set up in your home directory at ~/miniforge3 (this can also be a link to somewhere else, like /X2/SCRATCH/group/user/miniforge3). See [https://knowledge.ps.uci.edu/books/greenplanet-software/page/miniforge](https://knowledge.ps.uci.edu/books/greenplanet-software/page/miniforge) for details on preparing that. The miniforge/3 environment will use the base miniforge that is activated by loading the module with "module load miniforge/3". You will need to run "conda install jupyterhub jupyterlab batchspawner" to make sure the JupyterLab prerequisites are installed. ##### User JupyterHub If you want to keep your conda base environment clean, you can create a custom environment at ~/jupyterhub through either conda or pip. For example, starting with miniforge3's conda, do this to install a custom environment that is stored in your directory on /X2: ```bash mkdir /X2/SCRATCH/$(id -gn)/$USER/jupyterhub ln -s /X2/SCRATCH/$(id -gn)/$USER/jupyterhub ~/jupyterhub ml miniforge/3 conda create -y -p ~/jupyterhub batchspawner jupyterhub jupyterlab conda activate ~/jupyterhub conda install [list of needed packages for your notebook] pip install [list of any needed packages not in conda] ``` ##### Other Locations If you need to run in another arbitrary python environment, you can enter the name, location, and any environment modules you want to load manually in the "Advanced" tab. # X11 So one can run X11 (graphical) apps on the cluster and see the output on your local display, you may need to see one of the below links if you don't already have this configured. # X11 for Mac ##### MacOS 10.9 and later (11.2 “Big Sur” Working as of 4 March 2021): 1. Upgrade OS to latest available (not always necessary, but usually avoids bugs) 2. Upgrade to (or install) [XQuartz-2.8.5](https://www.xquartz.org/releases/XQuartz-2.8.5.html) 3. Log out/Log in to your Mac (may not be needed for upgrades from earlier 2.8.0 releases) 4. starting Mac terminal (or x2go, if installed) should trigger xquartz. 5. For OpenGL graphics to work, you may need to turn on indirect GLX by typing the following into your terminal: ``` defaults write org.xquartz.X11 enable_iglx -bool true ``` ##### Troubleshooting: - If `glxgears` gives the error: “couldn’t get an RGB, Double-buffered visual”, add the following to your GreenPlanet .bash\_profile or equivalent: ``` export __GLX_VENDOR_LIBRARY_NAME=mesa ``` - If Matlab windows keep blanking out, create a file "java.opts" in your GreenPlanet home folder with the following content: ``` -Dsun.java2d.xrender=false -Dsun.java2d.pmoffscreen=false ``` (If you start Matlab from a different folder, the java.opts may need to be put there instead.) ##### Before 10.9, use [XQuartz](https://www.xquartz.org/) [2.7.11](https://dl.bintray.com/xquartz/downloads/XQuartz-2.7.11.dmg) - - Warning: Some ancient apps may require an even *older* version (2.7.8) of XQuartz due to incompatibilities with “El Capitan” and “Mojave” # X11 for Windows Here are a list of X Windows servers for Windows ##### X2Go Requires an X2Go server on the target server. Creates a full remote desktop connection.| **Compiler/Version** | **To use** |
| gnu/4.8.5 gnu/7.2.0 gnu/7.3.0 gnu/8.1.0 gnu/8.2.0 gnu/8.3.0 gnu/9.1.0 | ml gnu/4.8.5 ml gnu/7.2.0 ml gnu/7.3.0 ml gnu/8.1.0 ml gnu/8.2.0 ml gnu/8.3.0 ml gnu/9.1.0 |
| intel/2017.2 intel/2017.4 intel/2017.5 intel/2018.0 intel/2018.1 intel/2018.2 intel/2018.3 | ml intel/2017.2 ml intel/2017.4 ml intel/2017.5 ml intel/2018.0 ml intel/2018.1 ml intel/2018.2 ml intel/2018.3 |
| intel/2018.0 + mkl intel/2018.1 + mkl intel/2018.2 + mkl intel/2018.3 + mkl | ml intel/2018.0 mkl ml intel/2018.1 mkl ml intel/2018.2 mkl ml intel/2018.3 mkl |
| **OpenMPI Version** | **Use GNU Compiler** | **Use Intel Compiler** |
| openmpi/1.10.7 | ml gnu/7.2.0 openmpi/1.10.7 | ml intel/2018.0 openmpi/1.10.7 |
| openmpi/2.1.1 | ml intel/2017.2 openmpi/2.1.1 | |
| openmpi/2.1.2 | ml gnu/7.2.0 openmpi/2.1.2 | ml intel/2017.4 openmpi/2.1.2 |
| openmpi/3.0.1a | ml gnu/7.2.0 openmpi/3.0.1a | ml intel/2018.0 openmpi/3.0.1a |
| openmp/3.0.1 | ml gnu/7.3.0 openmpi/3.0.1 | ml intel/2018.3 openmpi/3.0.1 |
| openmpi/3.1.1 | ml gnu/8.2.0 openmpi/3.1.1 | ml intel/2018.2 openmpi/3.1.1 ml intel/2018.3 openmpi/3.1.1 |
| openmpi/3.1.2 openmpi/3.1.2-slim | ml gnu/8.2.0 openmpi/3.1.2 ml gnu/8.2.0 openmpi/3.1.2-slim | ml intel/2018.3 openmpi/3.1.2 ml intel/2018.3 openmpi/3.1.2-slim |
Below is a table of all custom compiled software available on the new side of cluster (gplogin2/3) via lmod.
If compiling an app from source see Compiler Variants.
If compiling an MPI app from source see OpenMPI Variants.
This serves as a quick reference on how to load any package based on its prerequisite compiler if any.
One can generate a list of all available custom installed software via ml spider
Then ml spider <package> to get details on all variants of a given package.
The later is how the above table was generated!