Launch Your First Instance

This guide walks you through launching a GPU instance on NeuroHash.

Prerequisites

Before launching, make sure you have:

• A verified NeuroHash account
• ETH in your wallet on Robinhood Chain
• At least one SSH key added

info

If you're missing any of these, you'll see a yellow warning banner on your dashboard.

Step 1: Open the Launch Modal

From your Dashboard, click the Launch Instance button.

Step 2: Instance Details

Enter a name for your instance (e.g., "training-run-01").

Click Next to proceed.

Step 3: Configuration

Select Instance Type

Choose from available GPU configurations:

Instance Type	GPU	vCPUs	Memory	Storage	Price
gpu_1x_a10	1x A10	30	200 GiB	1.4 TB	~$0.75/hr
gpu_1x_a100	1x A100	30	200 GiB	1.4 TB	~$1.29/hr
gpu_8x_a100	8x A100	240	1800 GiB	21 TB	~$10.32/hr
gpu_1x_h100	1x H100	26	200 GiB	1 TB	~$2.69/hr
gpu_8x_h100	8x H100	208	1800 GiB	25 TB	~$21.52/hr

Prices shown are approximate USD equivalents. Actual charges are in ETH based on live exchange rates.

Select Region

Choose a region with available capacity. Only regions with GPUs currently available will be selectable.

Click Next to proceed.

Step 4: SSH Keys

Select one or more SSH keys to attach to the instance. At least one key is required.

Step 5: Launch

Click Launch Instance. The system will:

1. Verify your wallet has sufficient ETH balance
2. Deduct the first hour's cost from your wallet
3. Provision the GPU instance
4. Display the instance in your dashboard

Connecting to Your Instance

Once the instance status shows running, connect via SSH:

ssh ubuntu@<instance-ip>

The IP address is displayed in your dashboard's instance list.

Pre-installed Software

Every instance comes with:

• CUDA toolkit
• PyTorch (latest stable)
• TensorFlow
• Python 3.x
• JupyterLab
• Common ML libraries (numpy, pandas, scikit-learn, etc.)

Monitoring Usage

• Your instance's hourly cost is shown in the dashboard
• Billing is charged every hour from your wallet
• If your wallet balance runs out, the instance will be automatically terminated

Terminating an Instance

1. In the dashboard, select the instance(s) you want to terminate
2. Click Terminate
3. Confirm the action

caution

Terminating an instance is irreversible. All data on the instance's local storage will be lost. Save your work before terminating.

Troubleshooting

Issue	Solution
"Insufficient funds"	Fund your wallet with more ETH
No regions available	All GPUs of that type are currently in use. Try a different instance type or check back later
Can't connect via SSH	Ensure you selected the correct SSH key and that your local key matches