New NVIDIA H100 GPUs Now Available in Gcore Luxembourg GenAI Cluster
- January 30, 2024
- 2 min read
We’re pleased to announce the addition of more than one hundred new NVIDIA H100 GPUs with InfiniBand interconnect to our generative AI (GenAI) cluster in Luxembourg. These are the most advanced NVIDIA GPUs currently available. The new GPUs power Gcore Virtual Instances and Gcore Bare Metal Servers and are available with a six-month commitment. Read on to learn more about the H100, and if you’re interested in benefitting from Gcore AI GPU Infrastructure, please contact our sales team.
About Gcore H100 GPU Infrastructure
We offer the H100 within our AI GPU Infrastructure lineup, which includes NVIDIA A100 and L40S GPUs. The H100 provides impressive capabilities:
- Peak performance. With up to 7,916 teraFLOPS of performance, the NVIDIA H100 GPU delivers unparalleled processing power, making it ideal for training advanced generative AI and large language models (LLMs.)
- Ultra-fast networking. With InfiniBand interconnect, you can experience up to 3,200 Gbps of data transfer speed—the key to handling complex AI operations such as training LLMs, high-performance computing, and low-latency embedded I/O applications.
- Variety of configurations. We offer Virtual Instances with multiple GPU configurations for different AI/ML workloads. Bare Metal Servers are equipped with 8x H100 GPUs, 2x Intel Xeon 8468 CPUs, 2 TB RAM, 8x 3.84 TB NVMe SSD, and 3,200 Gbps InfiniBand.
- Scalability. With the new H100 and support for multi-GPU clusters, our infrastructure is highly adaptable to any ML workflow. We’re ready to handle any load and peak traffic spikes to maintain the level of performance required for your AI/ML applications.
- Kubernetes GPU worker nodes support. In addition to Virtual Instances and Bare Metal servers, NVIDIA H100 GPUs power Gcore Managed Kubernetes worker nodes. Run your containerized AI/ML workloads using the best GPU on the market.
- MLOps platform support. Gcore AI Infrastructure GPUs, including the H100, integrate seamlessly with MLOps platforms like UbiOps, streamlining ML workflows from data preparation to deployment.
How the H100 Stacks Up Against NVIDIA’s Other Options
To learn more about the H100’s capabilities and how they differ from other powerful GPUs, check out our article comparing the H100 to the NVIDIA A100, L40S, and H200 GPUs.
Conclusion
As a leading cloud AI service provider, we’re constantly evolving to offer the best technology on the market. Our new offering of H100 GPUs with InfiniBand upgrades reaffirms our commitment to providing world-class AI/ML development resources.
To try Gcore AI Infrastructure GPUs for yourself, fill out this form, and our sales team will be in touch.
Related articles
Earlier this year we brought NVIDIA Dynamo to Gcore — one-click disaggregated inference that delivered up to 6× higher GPU throughput and 2× lower latency inside a deployment, by separating prefill and decode and routing each request to the
For many startups, infrastructure decisions are mostly about performance, pricing, and developer experience. For Melious AI, they are also about trust.Melious AI is a German startup building a European AI platform around privacy, transparen
As enterprises move AI from experimentation into production, they face a new infrastructure challenge. AI applications, models, and data are no longer confined to a single cloud or data center. Instead, they are distributed across multiple
Subscribe to our newsletter
Get the latest industry trends, exclusive insights, and Gcore updates delivered straight to your inbox.