Working With Spaces#

This guide explains how to work with data stored in spaces.

Upload Input Files to a Space#

Follow these steps to upload files to a space in the CONVERGE Horizon web interface:

  1. Go to the Spaces page and click on the name of the space to which you want to upload files.

  2. Upload all of your input files into one directory. This will serve as the Case Directory from which you run your job.

    • To upload an entire directory from your local file system, click Upload Directory.

    • To upload files one by one, create a new directory (Menu > New Directory) and then click Upload Files to select individual files to upload.

Warning

When uploading files and folders, do not include the following characters in file names or directory names: & < > " '. Jobs will fail if any input file paths include any of these characters.

Additionally, we recommend maintaining spaces with fewer than 200,000 files at each directory level. If the number of files in a given directory approaches this limit, you might notice slow performance when your browser is loading those files. If you can no longer access a space from the web interface because of slow performance, contact horizon@convergecfd.com with the space ID.

You can optionally compress your input files into a zip or tar file before uploading the case. This file must be named inputs.zip or inputs.tar.gz, and it must contain the case files in the root directory of the archive. Create a new directory in the space as described above and upload the inputs.zip or inputs.tar.gz file to that directory. After you create the job, CONVERGE Horizon will automatically extract the files on the job cluster before starting the simulation.

There are a few other ways to upload files to a space without using the web interface. If you have installed the CLI, you can upload input files with the space:upload command. If you are working in CONVERGE Studio 5, you can use the CONVERGE Horizon button after adding your CONVERGE Horizon account settings to CONVERGE Studio in Edit > Preferences > CONVERGE Horizon.

Edit Files in a Space#

Files stored in your spaces can be edited from the CONVERGE Horizon web interface if they have an ASCII data format, such as .in, .dat, or .out. To edit a file, click on the file name. This will open a text editor where you can make and save your changes.

If you edit an input file for a job that is currently running, your changes will be saved to the space but will have no effect on the job. This is because the input files are copied from the space to the cluster at the start of the job. As CONVERGE runs, it reads data from the copies of the input files on the cluster. If you need to edit files on the cluster while a simulation is running, you can SSH to the head node.

Download Output Files From a Space#

For jobs that run from a space, CONVERGE Horizon uploads the output files from the job cluster to the space after the simulation ends. The output directory in the space has the same path as the input directory, with the job ID appended to the directory name. An example is shown below.

If you need to view or download output files while a job is running, you must SSH to the cluster.

../_images/output_directory.png

You can download a single file from the space by clicking the Download File (single-download-button) button or download multiple files by selecting them and clicking the Download Selected Files (multiple-download-button) button.

Downloading multiple files will work only if your browser is configured to (1) allow multiple downloads and (2) save downloaded files to a designated folder. In Chrome, the required settings are:

  • Set Automatic downloads to Allow for horizon.convergecfd.com.

  • Turn off Ask where to save each file before downloading.

If you use a different browser, refer to the browser’s documentation for information about settings that control this behavior.

Depending on how many files you select to download, you may notice a delay of several seconds or minutes before the downloads begin.

If you have installed the command-line interface, it may be faster to download multiple files at the terminal with the space:download command. If you work exclusively in the web interface and you typically download all of the output files for every job, consider using the option to compress output files when you create a new job.

Extract Compressed Output Files#

The files uploaded to a space from the job cluster are uncompressed by default. To conserve storage space and reduce download times, you can select the option to compress output files when you create a new job. If you select this option, the following tar files will be uploaded to the output directory:

  • image_files-<job ID>.tar.zst – All image files

  • log_files-<job ID>.tar.zst – All log files

  • output_files-<job ID>.tar.zst – All log files and output files, including 3D post*.h5 files

  • restart_files-<job ID>.tar.zst – All restart files

  • text_files-<job ID>.tar.zst – Text file output only (includes input, echo, log, and *.out files)

These files are compressed using Zstandard compression. To extract the files on Linux, you must have the zstd package installed. This is installed on most Linux systems by default. You can extract the files with the following command:

tar -I zstd -xf <file name>.tar.zst

Starting with Windows 11 version 23H2, Zstandard is natively supported on Windows, meaning you can extract the files the same way you would for any other type of archive. To extract the files on earlier versions of Windows, we recommend installing a program such as 7-zip-zstd.