https://github.com/thmslmr/gputree

Inspect Remote GPUs Activity
https://github.com/thmslmr/gputree

cli gpu monitoring python

Last synced: about 12 hours ago
JSON representation

Inspect Remote GPUs Activity

• Host: GitHub
• URL: https://github.com/thmslmr/gputree
• Owner: thmslmr
• License: mit
• Created: 2020-02-01T12:04:10.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2023-02-08T01:19:46.000Z (over 1 year ago)
• Last Synced: 2024-10-07T12:09:53.069Z (about 1 month ago)
• Topics: cli, gpu, monitoring, python
• Language: Python
• Homepage:
• Size: 35.2 KB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# `$ gputree`

**Record GPU activity on distant machine thru `nvidia-smi` command.**

Installation

---

```bash

pip install gputree

```

Hosts configuration

---

**PATH**

If no host is provided through the command line interface, `gputree` will look for a configuration file. By default, it looks for `gputree/config.yml` file in `$XDG_CONFIG_HOME` or `~/.config` directory. You can define a custom path to find the gputree configuration file by setting the `$GPUTREE_CONFIG_FILE` environement variable.

```bash

export GPUTREE_CONFIG_FILE=/path/to/gputree/config.yml`

```

**CONTENT**

In the configuration file, list your default hosts to look at under the `hosts` key. You can specify hosts in two ways :

1. Define a host with a unique name as key, and set its address with a user name as follows:

```yml

my-gpu:

    hostname: 0.0.0.0

    user: my.username

```

2. Refer to a host defined in your `~/.ssh/config` file if exists:

```yml

from-ssh-config:

    - gpu-1

    - gpu-2

```

**EXAMPLE**

Full `gputree` configuration file example:

```yml

hosts:

  my-gpu:

    hostname: 0.0.0.0

    user: my.username

  from-ssh-config:

    - gpu-1

    - gpu-2

```

License

---

This project is under [MIT License](./LICENSE)