AGAVE Access

This document provides useful references and step-by-step guidelines for accessing and utilizing the AGAVE supercomputer.

Useful References

Here are some useful references for information on AGAVE:

Prerequisites

Before accessing AGAVE, ensure the following prerequisites are met:

  1. VPN Connection: - Run Cisco Secure Client and login to the ASU network. - Under the connect field, enter sslvpn.asu.edu. - Click connect, and enter your ASURITE credentials, including the Duo Mobile Code for dual-factor authentication.

AGAVE Login Steps

Using SSH to access AGAVE supercomputer:

  1. SSH Login: - Run the following command to log in:

    ssh -Y ASURITE@agave.asu.edu

    • Replace ASURITE with your ASURITE username.

    • Enter your ASURITE password, ensuring Cisco Secure Client is running.

  2. Interactive Node: - Once logged in, switch from the login node:

    interactive

  3. Activate Conda Environment: - Activate your conda environment:

    source /home/ASURITE/anaconda3/bin/activate

Using Web Browser to access AGAVE Virtual Desktop:

  1. Open AGAVE Portal: - Click here to open the AGAVE Web Portal: AGAVE Web Portal. - Navigate to Agave Web Portal link on the ASU Research Computing website.

    AGAVE Web Portal

  2. Start Interactive Session: - Click Get started with the Agave Web Portal and navigate to My Interactive Sessions. - Select Agave Virtual Desktop under Interactive Apps. - Fill out the fields Partition, QOS, Cores, Session Wall Time, and click Launch. - Once resources are ready, click the Launch Agave Virtual Desktop button.

Loading Modules

  1. Check Available Modules: - Run the following to list all available modules:

    $ module -l avail

  2. Load Modules: - Load a specific module:

    $ module load module_name/version_number

  3. View Loaded Modules: - Check currently loaded modules:

    $ module list

  4. Unload Modules: - Unload a specific module:

    $ module unload module_name/version_number

    • Unload all modules:

      $ module purge

Partitions at AGAVE

Before submitting jobs, understand the partitions available:

  1. Check Partition Status: - For a specific partition:

    $ sinfo -p partition_name

    • For all partitions:

      $ sinfo -a

  2. Monitor Jobs: - Watch a specific job:

    $ squeue -j job_id

    • Watch all your jobs:

      $ watch 'squeue -u ASURITE'

Running pattern_sim through CrystFEL

Ensure you’re logged into AGAVE and have loaded the crystfel module.

Access and Execution:

  • Access the script at /home/ASURITE/Development/run_pattern_sim.sh.

  • Run pattern_sim through CrystFEL using run_pattern_sim.sh.

Arguments:

  • RUN: Run name (e.g., sim_run1).

  • GEOM_FILE: Path to the geometry file.

  • PDB_FILE: Path to the PDB file.

  • INTENSITY_FILE: Path to the intensity file.

  • TASKS: Number of tasks.

  • PARTITION: Partition name.

  • QOS: Quality of service.

  • TIME: Time limit.

  • TAG: Job tag.

Run the Script:

$ ./run_pattern_sim.sh RUN GEOM_FILE PDB_FILE INTENSITY_FILE TASKS PARTITION QOS TIME TAG

Indexing Images indexamajig through CrystFEL

Ensure you’re logged into AGAVE and have loaded the crystfel module.

Preparation:

  1. Verify .h5 files from pattern_sim.

  2. Create a list of images to index:

    $ ls /path/to/h5/files/*.h5 > my_images.list

  3. Load necessary modules (ccp4 and crystfel).

Indexing:

  • Index images using indexamajig:

    $ indexamajig -i my_images.list -o output.stream -g geom_file.geom --additional-options

Refer to SLURM Scripts <slurm_script> for more information on submitting jobs.

Note

Adjust arguments as necessary and refer to CrystFEL documentation for detailed usage.