GPU Rental Guide
Complete guide to renting GPU instances on SynpixCloud
Overview
This guide walks you through the complete process of renting a GPU instance on SynpixCloud, from selecting your GPU to connecting to your running instance.
Step 1: Select GPU
Browse our GPU Marketplace to find the GPU that fits your needs.
Selection Options
- GPU Type - Choose from RTX 4090, A100, H100, and more
- Quantity - Select the number of GPUs you need
- CPU & Memory - Configure CPU cores and RAM allocation
- Rental Duration - Choose hourly, daily, or monthly billing
CPU and memory specifications can be adjusted after instance creation.
Step 2: Configure Instance
Image Selection
Choose a pre-configured environment based on your workflow:
| Image | Included | Best For |
|---|---|---|
| PyTorch | PyTorch, CUDA, cuDNN | Deep learning training |
| TensorFlow | TensorFlow, CUDA | ML model development |
| Ubuntu Base | Ubuntu 22.04, CUDA drivers | Custom environments |
| ComfyUI | ComfyUI, Python, dependencies | AI image generation |
Disk Configuration
- System Disk - Base storage for OS and software
- Data Disk - Additional storage for datasets and models
- Disk Retention - Choose whether to preserve data when instance is stopped
Select "Retain Disk" if you want to preserve your data between sessions. Enabling disk retention may incur additional storage fees.
Step 3: Review & Create
Before creating your instance, confirm:
- GPU Configuration - Type, quantity, specifications
- Billing - Hourly rate, billing cycle, total cost
- Auto-Renewal - Enable for continuous operation
- Storage - Disk capacity and retention settings
- Network - Bandwidth allocation
Payment
The full rental amount is deducted from your account balance at creation time. Ensure sufficient balance before proceeding.
Click Create Instance to launch your GPU.
Step 4: Connect to Instance
Once created, your instance will be ready within 1-2 minutes. The status will change to Running.
Get Connection Info
- Find your instance in the dashboard
- Click Connection Details to view:
- SSH command and port
- Root password
- Jupyter notebook URL (if enabled)
- Mapped ports
SSH Connection
ssh -p <port> root@<host>Enter the provided password when prompted.
Jupyter Access
If Jupyter is enabled, access it via the provided URL in your browser.
Instance Management
Available Actions
From your dashboard, you can:
| Action | Description |
|---|---|
| Start/Stop | Control instance power state |
| Restart | Reboot the instance |
| Upgrade | Increase CPU/memory allocation |
| Extend | Add rental time to your instance |
| Reset | Reinstall the system image |
| Terminate | Permanently delete the instance |
Instance Status
| Status | Description |
|---|---|
| Provisioning | Instance is being created |
| Running | Instance is active and accessible |
| Stopped | Instance is powered off |
| Stopping | Instance is shutting down |
| Expired | Rental period ended, awaiting action |
Tips for Cost Optimization
- Stop when idle - Stop instances when not in use (if disk retention is enabled)
- Choose daily/monthly - Get 2-5% discounts on longer commitments
- Right-size your instance - Don't over-provision CPU/memory
- Enable auto-renewal - Avoid service interruption for long-running jobs
Support
Need help? Contact us at support@synpixcloud.com