Skip to main content

How to run GPU workloads on E2E Networks

Let's discover how to launch and run GPU workloads on E2E Networks.

Getting Started​

E2E Networks is an India-focused cloud provider with GPU options such as H100 and H200. It is a useful option when you want lower-latency access in India and flexible AI/ML infrastructure.

Get started by creating an account at https://myaccount.e2enetworks.com/. You can launch GPU workloads from the TIR AI/ML platform.

What you'll need​

Launch a GPU instance​

  1. Go to AI cloud (TIR). Select the base image (for example, Python 3.10 image).

  2. Give the instance a name (for example, example-node).

  3. Choose the GPU type required for your workload (for example, H200 for large inference/training jobs).

  4. Configure storage, SSH keys, and static IP (if needed), then launch the instance.

  5. Connect to the instance over SSH.

  6. Update system packages:

sudo apt update && sudo apt upgrade -y

Verify GPU access​

After connecting, run:

nvidia-smi

If everything is configured correctly, you should see your GPU model, driver version, and CUDA details.

Cost optimization tips​

  • Pick the smallest GPU that meets your latency/throughput target.
  • Stop instances when not in use to reduce cost.
  • Use larger GPU tiers only when your model size or concurrency requires them.
  • Track runtime and storage usage together; storage can add up over time.

Screenshots​

Add screenshots from the E2E console and terminal output (for example, instance config and nvidia-smi) to mirror the style used in other provider docs.