We're thrilled to announce, that Open Telekom Cloud supports now the NVIDIA H100 GPUs. The NVIDIA H100 Tensor Core GPU is the latest in high-performance computing and AI acceleration. Designed to deliver unparalleled performance for AI, machine learning, and high-performance computing (HPC) workloads, the H100 GPUs provide advanced features for data-intensive operations.
Key features:
- Powered by the NVIDIA Hopper architecture, delivering up to 30x performance improvements over previous-generation GPUs for AI inference and training workloads.
- FP32, FP64, and TensorFloat-32 support for a range of precision requirements.
- Higher L2 Cache and enhanced bandwidth for higher throughput.
- Transformational support for large language models (LLMs) and generative AI tasks.
- Power-efficient design to reduce data center carbon footprints.
- Pay-as-you-go and reserved options tailored to meet budget needs.
- Great support for frameworks like TensorFlow, PyTorch, and NVIDIA CUDA-X AI libraries for seamless integration.
The NVIDIA H100 GPUs are now available as ECS p5s flavor family in the EU-DE region of the Open Telekom Cloud.
Be the first maximizing your compute power with NVIDIA H100!