NVIDIA H200 NVL
GPU architecture NVIDIA Hopper
GPU accelerator optimized for generative AI and HPC computing
14592 NVIDIA CUDA cores for large-scale parallel computing
456 NVIDIA Tensor cores for AI model training and inference
141 GB of HBM3e memory with ECC - huge space for the largest AI models
Memory bandwidth up to 4,8 TB/s
Interface PCIe 5.0 x16 providing very high data throughput
Passive cooling designed for servers and GPU clusters
Maximum power consumption: 600 W
Free shipping from €300
Promocja cenowa na model HDR-15-5
Product intended for professional use only
NVIDIA H200 NVL
Description
NVIDIA H200 NVL GPU accelerator for generative AI and the largest LLM models
NVIDIA H200 Tensor Core GPU is one of the latest and most advanced compute accelerators designed for artificial intelligence infrastructure and high performance computing (HPC). Based on the NVIDIA Hopper architecture, the chip delivers breakthrough performance and massive memory bandwidth to support the most demanding AI workloads.
The H200 NVLGPU is designed to train and run very large artificial intelligence models, including generative AI systems used in enterprises and data centres. With a new generation of HBM3e memory, the accelerator offers significantly higher bandwidth and memory capacity than previous generations of GPUs.
NVIDIA Hopper architecture and HBM3e memory
The accelerator uses the NVIDIA Hopper architecture, which is designed to maximise the acceleration of AI computing. The GPU is equipped with 14592 CUDA cores and 456 Tensor cores to accelerate mathematical operations used in deep learning.
The H200's most significant innovation is the use of 141 GB of HBM3e memory with ECC error correction and a bandwidth of up to 4.8 TB/s. This enables the GPU to support the largest artificial intelligence models and huge data sets used in modern AI systems.
CUDA cores
Tensor cores
of HBM3e memory
memory bandwidth
GPU for generative artificial intelligence
The NVIDIA H200 NVL is designed as an accelerator for the most advanced generative artificial intelligence systems. The GPU enables massive AI models to be trained and rapidly deployed in production environments.
training and inference of LLM models
VLM and systems combining text, image and video
AI using enterprise knowledge
AI-as-a-Service platforms
big data and AI analytics
academic and research computing
Accelerator for next-generation data centres
The NVIDIA H200 NVL GPU is designed to run in modern server platforms and GPU clusters used in data centres. The PCIe 5.0 x16 interface provides ultra-high bandwidth communication between the GPU and the server system.
With passive cooling and a maximum power consumption of 600 W, the accelerator is designed for professional AI infrastructure and HPC systems serving the most demanding workloads.
GPUs for AI infrastructure and supercomputing
The increasing complexity of artificial intelligence models is driving the need for massive computing power and very high memory bandwidth. The NVIDIA H200 NVL addresses these needs, enabling scalable AI platforms used in enterprises, data centres and research institutes.
The accelerator is a key component of the infrastructure used to train the largest artificial intelligence models and conduct advanced scientific computing.
Technical Specification
| Specification | H200 SXM | H200 NVL |
|---|---|---|
| FP64 | 34 TFLOPS | 30 TFLOPS |
| FP64 Tensor Core | 67 TFLOPS | 60 TFLOPS |
| FP32 | 67 TFLOPS | 60 TFLOPS |
| TF32 Tensor Core² | 989 TFLOPS | 835 TFLOPS |
| BFLOAT16 Tensor Core² | 1,979 TFLOPS | 1,671 TFLOPS |
| FP16 Tensor Core² | 1,979 TFLOPS | 1,671 TFLOPS |
| FP8 Tensor Core² | 3,958 TFLOPS | 3,341 TFLOPS |
| INT8 Tensor Core² | 3,958 TFLOPS | 3,341 TFLOPS |
| GPU Memory | 141 GB | 141 GB |
| GPU Memory Bandwidth | 4.8 TB/s | 4.8 TB/s |
| Decoders | 7 NVDEC 7 JPEG |
7 NVDEC 7 JPEG |
| Confidential Computing | Supported | Supported |
| Max Thermal Design Power (TDP) | Up to 700W (configurable) | Up to 600W (configurable) |
| Multi-Instance GPUs | Up to 7 MIGs @ 18GB each | Up to 7 MIGs @ 16.5GB each |
| Form Factor | SXM | PCIe Dual-slot air-cooled |
| Interconnect | NVIDIA NVLink: 900GB/s PCIe Gen5: 128GB/s |
2- or 4-way NVIDIA NVLink bridge: 900GB/s per GPU PCIe Gen5: 128GB/s |
| Server Options | NVIDIA HGX H200 partner and NVIDIA-Certified Systems with 4 or 8 GPUs | NVIDIA MGX H200 NVL partner and NVIDIA-Certified Systems with up to 8 GPUs |
| NVIDIA AI Enterprise Add-on | Included | Included |
¹ Preliminary specifications. May be subject to change.
² With sparsity.
