NWS
GPUs
GPU power for AI, machine learning and high-performance computing
Our available GPUs
NVIDIA A10
- For small models
NVIDIA A40
- For medium-sized models
NVIDIA RTX Pro 6000
- For large models
| wdt_ID | Merkmal | NVIDIA A10 | NVIDIA A40 | NVIDIA RTX Pro 6000 |
|---|---|---|---|---|
| 14 | Architektur | Ampere-Architektur | Ampere-Architektur | Blackwell Architektur |
| 15 | Speicher | 24GB GDDR6 VRAM | 48GB GDDR6 ECC VRAM | 96GB GDDR7 ECC VRAM |
| 16 | Speicherbandbreite | 600 GB/s | 798 GB/s | 1597 GB/s |
| 17 | Rechenleistung | FP32 31 teraFLOPS | FP32 37 teraFLOPS | FP32 120 teraFLOPS, FP4 AI Compute (Peak) 4 petaFLOPS |
| 18 | Einsatzbereiche | AI-Inference, Datenanalyse | Simulationen, Deep Learning & AI-Inference | komplexe Simulationen, Deep Learning, AI-Finetuning & AI-Inference |
| 19 | AI-Modelle | z.B. GPT-OSS-20B oder Qwen3-8B | z.B. Ministral-3-14B oder Qwen3-14B | z.B. GPT-OSS-120B oder Qwen3.5-27B |
| 21 | MIG (Multi-Instance-GPU) | X | X | ✓ |
Your advantages
Maximum performance for demanding workloads
High availability
Stable infrastructure for critical applications.
Optimized performance
Modern hardware with high computing power and scalable resources.
Security and data protection
Our data centers in Germany are ISO-certified and we operate in compliance with the GDPR.
Integration into existing workflows
Easy integration into existing cloud or Kubernetes environments.
Create an NWS ID and get started right away!
Typical use cases
Artificial Intelligence & Machine Learning
High Performance Computing
Data Science & Big Data
Rendering & Visualization
Next steps
Step-by-step instructions for our AI service.
Find out more about our team, our values and our vision.
Exciting articles, helpful tips and up-to-date information.
Comprehensive technical documentation for optimal use.
FAQ
2
3
When is the use of GPUs worthwhile?
The use of GPUs is particularly worthwhile for computationally intensive tasks. These include machine learning, large data analyses, simulations and rendering processes. GPUs can process these workloads significantly faster than traditional CPU systems.
2
3
Which applications particularly benefit from GPUs?
Typical applications are
- Training of AI and machine learning models
- Data analysis and big data processing
- 3D rendering and visualization
- Scientific simulations
- Video processing
2
3
Can GPU resources be scaled flexibly?
Yes, the GPU resources can be scaled flexibly. Depending on requirements, additional GPUs can be provided or workloads can be distributed across several GPUs.
2
3
Are GPUs also useful for small projects?
Even smaller projects can benefit from GPU acceleration, for example when training initial machine learning models or processing large data sets. GPUs can be used as required thanks to flexible provisioning models.
2
3
How is GPU performance measured?
The performance of a GPU is determined by various factors, including
- Architecture of the GPU
- Memory size and memory bandwidth
- Computing power
2
3
Who can I contact if I have problems?
Please contact us by e-mail at nws@netways.de or via our contact form.