Stellar.AI
  • Overview
    • Overview
    • Architecture
  • Rendering
    • Getting started
    • Top-up your account
    • Setup add-on
    • Submitting your first render
    • Start your render
      • Transfer manager (Uploads and downloads)
      • Managing your render
      • Terminate rental
  • VPN
    • Android
    • iOS
  • AI and ML
  • Resize Docker Storage
  • Nodes
    • Installation
    • Update/Uninstall
    • Idle jobs
    • WSL
    • Verification
  • Services
    • CPU/GPU rental
  • Cryptocurrency
    • Monetary Policy
    • Staking
  • Roadmap
  • API Reference
    • accounts
      • /accounts
      • /accounts
      • /accounts/balance
    • network
      • /network
    • nodes
      • /nodes/{id}
    • services
      • /services
      • /services/mr
      • /services/mr
      • /services/render
      • /services/render
      • /services/{uuid}/info
      • /services/{uuid}/stop
Powered by GitBook
On this page

Nodes

PreviousResize Docker StorageNextInstallation

Last updated 5 months ago

In general the node, aka st-node is a computer with available of CPUs, GPUs, memory and disk space.

sta Nodes are the foundation of our compute marketplace, providing the necessary computational power to meet the demands of heavy tasks. These computers are equipped with a blend of CPUs, GPUs, memory, and disk space that allows them to handle heavy and distributed workloads with ease. The nodes are connected to the staSpace network, which enables them to seamlessly work together to deliver optimal performance and efficiency...

Software requirements

The following OS and system architectures are supported:

System requirements

Staking and VPN

Minimal system requirements for staking and VPN nodes

1 CPU 1 Gb RAM 10 Gb free disk space Static IPv4 address

GPU nodes

Minimal system requirements for GPU nodes

2 CPU cores 16 Gb RAM 100 Gb free disk space Static IPv4 address

NVIDIA or AMD GPU.

Rendering

Rendering is an intensive process which utilizes not just GPU's but all system resources in the processing and rendering frames.

If you wish your node to support rendering services there are more specific requirements you need to pay attention to.

RAM requirements

Depending on the VRAM and number of GPU's, you will need to ensure you scale your RAM accordingly.

The rule is:

VRAM x GPU’s x 2 for RAM

For example a node with 2 x RTX 4090 (24GB VRAM) would require minimum of 96GB RAM (24 x 2 x 2).

CPU requirements

The CPU plays a critical role in pre-processing of the render and compiling shaders as well as managing the workloads between the GPU's and therefore needs to have sufficient cores to support and scale with the GPU's.

The optimal CPU should have 8 cores per GPU.