Getting Started with TerraGPU

Quick Start (15 minutes)

Step 1: Create Account

Sign up at terragpu.com/signup

Verify your email and complete onboarding (organization, payment method)

Step 2: Set Up Network

Choose connectivity option:

  • TerraGPU VPN (Recommended): 5-minute setup, no AWS account needed
  • Customer VPN: Connect your existing VPN
  • Direct Connect: Dedicated 10Gbps+ connection

See: VPN Setup Guide

Step 3: Launch GPU Instance

Go to Console → Instances → Launch Instance

  1. Select GPU type (1x A100 recommended for testing)
  2. Choose container image (PyTorch, TensorFlow, or custom)
  3. Configure storage (500GB ephemeral included)
  4. Set name and options
  5. Review and launch

Instance ready in 2-3 minutes

Step 4: Connect & Run

Connect via SSH or Jupyter:

ssh ubuntu@10.100.5.23
nvidia-smi  # Check GPUs
python train.py  # Run your code

What You Get

  • NVIDIA A100/H100 GPUs with CUDA 12.2
  • Ubuntu 22.04 LTS with pre-installed ML frameworks
  • 500GB NVMe ephemeral storage per instance
  • Private networking via VPN
  • Per-second billing (1-minute minimum)

Next Steps

Need Help?

Email: support@terragpu.com

Response time: Within 24 hours