# H100 Series Tensor Technology

* **Tensor Core Technology**
  * Optimized for AI and deep learning with support for multiple precision formats (FP8, FP16, BF16, TF32, FP64)
  * Enables faster computations and higher throughput for AI workloads
* **HBM3 Memory**
  * Provides up to 3 TB/s of memory bandwidth
  * Enhances data transfer speeds and efficiency for processing large datasets
* **NVIDIA Grace CPU Interconnect**
  * Allows pairing with NVIDIA Grace CPU via NVLink for high-speed data transfer
  * Optimizes performance for AI and HPC workloads
* **Multi-Instance GPU (MIG)**
  * Enables partitioning of a single GPU into smaller instances for isolated access
  * Maximizes utilization and flexibility in shared computing environments
* **DPX Instructions**
  * Accelerates dynamic programming algorithms
  * Key for applications in genomics, quantum chemistry, and robotics
* **Transformer Engine**
  * Specifically designed to accelerate transformer models used in NLP and generative AI
  * Optimizes performance, reducing training and inference times
* **Security Features**
  * Includes secure boot and a hardware root of trust
  * Ensures a secure computing environment and protection against threats
* **Energy Efficiency**
  * Delivers exceptional performance per watt
  * Reduces operational costs and supports sustainability by minimizing carbon footprint
* **Scalability**
  * NVLink and NVSwitch technologies enable efficient scaling across multiple GPUs
  * Allows for the construction of powerful GPU clusters for demanding tasks
* **Compatibility with CUDA and AI Frameworks**
  * Fully supports NVIDIA's CUDA-X libraries and popular AI frameworks
  * Ensures easy deployment of applications and leverage of AI software advancements


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://tpuai.gitbook.io/tpu-ai-whitepaper/features/h100-series-tensor-technology.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.