> ## Documentation Index
> Fetch the complete documentation index at: https://docs.rightnowai.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Remote GPU Execution

> Run CUDA code on remote GPUs through SSH.

# Remote GPU Execution

Connect to remote GPUs on cloud providers or your own servers.

## Quick Start

<Frame>
  <img src="https://mintcdn.com/rightnowai/MyRQ6ILCdazu_XAM/images/remote-gpu-connection-dialog.webp?fit=max&auto=format&n=MyRQ6ILCdazu_XAM&q=85&s=d289b64b5ce61579a9efa9626eed354d" alt="Remote GPU Connection Dialog" width="1302" height="640" data-path="images/remote-gpu-connection-dialog.webp" />
</Frame>

<Steps>
  <Step title="Connect to Remote GPU">
    Press `Ctrl+Shift+G` or click "Remote GPU: Connect" in command palette.
  </Step>

  <Step title="Choose Connection Type">
    Select one:

    * **Quick Connect**: Paste your SSH command
    * **RunPod/Vast.ai/Lambda**: Select your provider
    * **Manual Setup**: Enter server details
  </Step>

  <Step title="Enter Connection Info">
    Provide your SSH details:

    * Host/IP address
    * Username
    * SSH key or password
  </Step>

  <Step title="Start Using Remote GPU">
    Once connected, your code runs on the remote GPU automatically.
  </Step>
</Steps>

## Supported Cloud Providers

### RunPod

* Copy SSH command from RunPod dashboard
* Paste in Quick Connect
* RightNow handles the rest

### Vast.ai

* Get instance SSH details
* Use Quick Connect or Manual Setup
* Supports custom ports

### Lambda Labs

* Use instance IP address
* Username: ubuntu
* Add your SSH key

### Paperspace

* Get SSH endpoint from console
* Username: paperspace
* Connect via Manual Setup

## How to Use

### Open Remote Terminal

<Frame>
  <img src="https://mintcdn.com/rightnowai/MyRQ6ILCdazu_XAM/images/ssh-terminal-remote.webp?fit=max&auto=format&n=MyRQ6ILCdazu_XAM&q=85&s=91a854ed57050d006bc46abcbc07e0f1" alt="SSH Terminal Remote" width="1907" height="974" data-path="images/ssh-terminal-remote.webp" />
</Frame>

Press `Ctrl+Shift+T` to open SSH terminal to your remote GPU.

### Run CUDA Code

1. Open your .cu file
2. Click Build button
3. Code compiles and runs on remote GPU

### Switch GPUs

Click GPU name in status bar to switch between local and remote.

## Connection Status

<Frame>
  <img src="https://mintcdn.com/rightnowai/MyRQ6ILCdazu_XAM/images/remote-gpu-status-bar.webp?fit=max&auto=format&n=MyRQ6ILCdazu_XAM&q=85&s=4ffe7904ae6df48242f6e22cc13779ce" alt="Remote GPU Status Bar" width="717" height="344" data-path="images/remote-gpu-status-bar.webp" />
</Frame>

Look at the status bar:

* **Connected**: Shows remote GPU name
* **Disconnected**: Shows local GPU
* **Connecting**: Shows progress

## Tips

<Check>
  **Best Practices**

  * Use SSH keys instead of passwords
  * Save connections for quick access
  * Check GPU availability with terminal
  * Monitor GPU usage in status bar
</Check>

<Warning>
  **Troubleshooting**

  * Can't connect? Check firewall and SSH settings
  * GPU not detected? Run `nvidia-smi` in terminal
  * Slow connection? Check network speed
</Warning>

## Screenshots Needed

1. Connection dialog with provider options
2. Remote GPU shown in status bar
3. SSH terminal to remote GPU

## Next Steps

* [GPU Emulation](/gpu-emulation) - Test without hardware
* [Benchmarking](/benchmarking) - Compare performance
* [Core Features](/core-features) - AI assistance