Core Features of GPUiq
Explore the essential features of GPUiq that enable efficient GPU rentals for various workloads like AI and rendering.
Overview
GPUiq provides powerful GPU rentals optimized for AI training, machine learning inference, 3D rendering, and cryptocurrency mining. You access high-performance hardware at affordable prices, with up to 70% cost savings compared to traditional cloud providers. Key features include diverse GPU types, automatic switching for resource optimization, snapshot functionality for data preservation, and flexible pricing models.
Start by selecting a GPU type that matches your workload. GPUiq automatically scales resources to minimize costs and downtime.
Available GPU Types
GPUiq offers a range of GPUs with detailed performance specifications. Choose based on your needs for VRAM, compute power, and tensor performance.
| GPU Model | VRAM | FP32 TFLOPS | Tensor Cores | Ideal For |
|---|---|---|---|---|
| NVIDIA A100 | 80GB | 19.5 | 432 | AI training, large models |
| NVIDIA H100 | 141GB | 67 | 528 | High-end ML, HPC |
| RTX 4090 | 24GB | 82.6 | 304 | Rendering, gaming sims |
| A6000 | 48GB | 38.7 | 336 | Balanced ML/rendering |
High VRAM Options
Select A100 or H100 for memory-intensive tasks like training LLMs.
Cost-Effective Choices
Use RTX 4090 for rendering workloads with excellent price/performance.
Automatic Switching
Automatic switching dynamically allocates the optimal GPU based on your workload demands. GPUiq monitors usage and switches seamlessly to prevent overprovisioning.
Enable Auto-Switching
In your dashboard, toggle autoSwitch to true.
Define Rules
Set thresholds: minVRAM: "24GB", maxTFLOPS: 50.
Monitor Transitions
View logs for switch events.
Snapshot Functionality
Snapshots capture your GPU instance state, including data and models, for quick restoration. Preserve progress without full recreation.
curl -X POST https://api.example.com/v1/snapshots \
-H "Authorization: Bearer YOUR_TOKEN" \
-d '{
"instanceId": "gpu-12345",
"name": "model-training-v1"
}'
curl -X POST https://api.example.com/v1/instances \
-H "Authorization: Bearer YOUR_TOKEN" \
-d '{
"snapshotId": "snap-67890",
"gpuType": "A100"
}'
Cost-Saving Mechanisms
GPUiq delivers up to 70% savings through on-demand pricing, reserved instances, and spot markets. Pay only for active usage with automatic shutdowns.
# CLI: List pricing
gpuiq pricing list --region us-east
// JS SDK
const pricing = await gpuiq.pricing.list({ region: 'us-east' });
console.log(pricing.spot.a100); // $0.50/hour
| Model | Hourly Rate (A100) | Savings |
|---|---|---|
| On-Demand | $1.20 | Baseline |
| Reserved (1yr) | $0.72 | 40% |
| Spot | $0.36 | 70% |
Combine spot instances with snapshots for maximum savings on interruptible workloads.
Last updated today