Renting GPU Compute

Renting GPU Compute on FluxEdge

FluxEdge offers three types of machines for rent: Dedicated GPU machines from community providers with graphics cards for GPU-intensive tasks, Dedicated CPU machines for CPU-only workloads, and Premium machines featuring enterprise-grade GPUs (H100, A100, Blackwell) from trusted partners like Hyperstack/NVIDIA. This guide walks through the complete rental process.

Machine Types

Finding the Right Machine

The Rent Machine page provides powerful filtering and sorting to find the optimal machine for your workload.

• •Basic filters: Operating system, region (geographic location), price range
• •Advanced filters: GPU VRAM, CPU core count, network speed, specific GPU model (e.g., RTX 4090), storage type (NVMe)
• •Key metrics: Price per hour, Score/$/hour (cost-performance ratio), benchmark score, reliability percentage, uptime

Renting a Dedicated Machine

1. 1

   Select your machine

   Browse the available machines list. Review specs, pricing, reliability score, and location.

2. 2

   Click "Rent"

   Click the Rent button at the top right of the machine card.

3. 3

   Confirm selection

   Review the confirmation popup with pricing details and confirm.

4. 4

   Wait for provisioning

   The machine enters a provisioning state. This typically takes under 10 minutes (maximum 30 minutes). The stages are: Provisioning → Installing FluxCore → Configuring Cluster → Installing GPU Operator.

5. 5

   Deploy your workload

   Once provisioning completes, deploy your application via Quick Launch or Custom Deployment.

Renting a Premium Machine

Premium machines require a slightly different workflow since they are sourced from enterprise partners.

1. 1

   Select GPU type and quantity

   Choose from H100, A100, H200, RTX A6000, or Blackwell GPUs. Select how many GPUs you need.

2. 2

   Choose hardware specs

   Configure CPU, RAM, and storage to match your workload requirements.

3. 3

   Select server region

   Choose a geographic region for your deployment. Latency-sensitive workloads benefit from closer regions.

4. 4

   Specify duration

   Set the number of machines and rental duration. Review the price per hour and any one-time startup fee.

5. 5

   Confirm and deploy

   Review total cost, confirm, and wait for provisioning to complete.

Machine Management

Once a machine is rented, the Machine Overview screen provides complete visibility into your rental.

• •Machine Info: Name (hash ID), type, uptime, reliability percentage, hardware specs, location
• •Projects: List of deployed workloads running on the machine. No limit on the number of projects.
• •Shell Access: Direct SSH/shell access to your pods. Use tmux for long-running sessions (shell may disconnect on inactivity).
• •Logs: Container logs viewable per container. Logs are retained for 10 days after deployment completion.
• •Actions: Stop Machine, Add New Project, Redeploy, view detailed machine Info