Build, Deploy, and Scale AI Solutions across the Enterprise
Intel is empowering developers to run end-to-end AI pipelines with Intel® Xeon® Scalable processors. From data preprocessing and modeling to production, Intel has the software, compute platform, and solution partnerships you need to accelerate the integration of AI everywhere.
See the performance difference for yourself.
AI Platform Overview
AI Software
Build and deploy everywhere with oneAPI-powered tools and framework optimizations to accelerate the AI journey.
Compute Platform
Run any AI workload on Intel Xeon processors optimized for end-to-end AI performance.
Solutions & Partnerships
Implement prebuilt solutions and access an extensive partner ecosystem and products for faster time to solution.
Run Any AI Workload with the New 4th Gen Intel® Xeon® Scalable Processor
Combining the flexibility of a general-purpose CPU with the performance of a deep learning accelerator, the new 4th gen Intel Xeon Scalable processor uses the industry's most popular open standard frameworks with little to no code changes.
Featuring Intel® Advanced Matrix Extensions (Intel® AMX) in Every Core for Deep Learning
The new 4th gen Intel Xeon Scalable processor features architecture improvements such as Intel® AMX that accelerate classical machine learning and deep learning workloads. Intel AMX allows most deep learning inference models and small and medium-sized deep learning training models to run on the same platform.
Intel AMX has two primary components: tiles and tiled matrix multiplication (TMUL). The tiles store large amounts of data in eight two-dimensional registers, each one kilobyte in size. TMUL is an accelerator engine attached to the tiles that contains instructions to compute larger matrices in a single operation.
Intel® AI Software Enables AI Everywhere Faster
Accelerating development with optimizations of the most popular industry libraries and open source tools, the Intel AI software suite unleashes the performance of Intel Xeon Scalable processors without code changes.
The Intel AI software suite has been validated on over 400 AI models and use cases to help ensure that you achieve out-of-the-box application performance.
Performance
Accelerate performance using any popular open source library and deep learning framework optimized for Intel Xeon Scalable processors
Productivity
Develop faster with a portable stack, the open, cross-architecture oneAPI programming model, and pretrained AI models.
Simplicity
Automate the development of high-performance pipelines and expand data science accessibility with Intel AI tools and kits.
Technology Innovators Achieved Impressive Results
These Intel partners are using the 4th gen Intel Xeon Scalable processor together with Intel AI software to build platforms and services that achieve their deep learning and general-purpose performance targets—all on a single universal platform.
Running Fujitsu's Digital Transformation Platform
4x BERT Inferencing Improvement with OpenVINO™ for Sentiment Analysis1
Fujitsu* chose Intel Xeon processors and Intel software to build the Fujitsu DXP (digital transformation platform) for use in the data center to the factory floor. Its AI Test Drive helps developers experiment with end-to-end AI pipeline technologies such as natural language processing (NLP) and computer vision.
Conversational AI That Gets It Right
Numenta* Delivers More Accurate Chatbot Responses with 62x NLP Performance2
Numenta applied advanced neuroscience research to achieve breakthrough performance across NLP and computer vision. Numenta and Intel technology led to breaking the 10 milliseconds latency barrier for BERT-Large and achieving 62x improvement.
Visualize the Difference with Graph Computing
1.6x Faster GNN Training Compared to NVIDIA* A1003
Katana Graph* uses the 4th gen Intel Xeon Scalable processor to build a platform that can easily scale out on multiple standard server nodes for massive data processing, analytics, and graphic neural networks to gain unprecedented levels of performance. This scale is enabling pharma companies to reduce costs and bring life-saving drugs to the market faster.
AI Made Easy with a Low Code or No Code Platform
Democratize AI for Everyone 2x Faster with Aible* and Intel4
Aible* uses Intel Xeon processors to build a single universal AI platform to accommodate any AI use case for its customers. For each project, Aible evaluates hundreds of model types across many hyperparameters to guarantee AI impact in 30 days.
Expect Amazing Deep Learning Training and Inferencing Performance Results
With a built-in bfloat16 and int8 deep learning matrix multiplication engine integrated into every core, Intel AMX delivers up to 10x higher generation-to-generation deep learning inferencing and training performance.5
Transfer Learning in Minutes
With industry standard frameworks and access to tens of thousands of pretrained models and tools, the Intel® AI developer platform featuring Intel AMX enables cost-effective transfer learning. In a comparison between the NVIDIA A100* and the 4th gen Intel Xeon Scalable processor, PyTorch* and TensorFlow* training is completed in less than four minutes.6
Continued Leading Performance for Machine Learning
Intel Xeon processors continue to excel with machine learning. The 4th gen Intel Xeon Scalable processor beats AMD* EPYC7 and NVIDIA* A1008 for machine learning training and inference in performance benchmarks.
Resources
MLOps Automation
End-to-end MLOps automation lets you easily run your AI workloads on any infrastructure with Intel® Tiber™ AI Studio. Bring your model and try the latest Intel Xeon processor performance with Intel Tiber AI Studio.
Intel® Developer Cloud
Develop and run workloads in Intel-optimized deployment environments with the latest Intel processors and performance-optimized software stacks.
OpenVINO™ Toolkit
An open source toolkit for optimizing and deploying deep learning inference solutions.
Product and Performance Information
Availability of accelerators varies depending on SKU. For additional product details, see Intel Product Specifications.
1. Fujitsu performance. See [P5] of the Performance Index for 4th Gen Intel Xeon Scalable processors. Results may vary.
2. Numenta performance. See [P6] of the Performance Index for 4th Gen Intel Xeon Scalable processors. Results may vary.
3. Katana Graph. See [P7] of the Performance Index for 4th Gen Intel Xeon Scalable processors. Results may vary.
4. Aible performance. See [P4] of the Performance Index for 4th Gen Intel Xeon Scalable processors. Results may vary.
5. Intel AMX deep learning training and inference. See [A17] of the Performance Index for 4th Gen Intel Xeon Scalable processors. Results may vary.
6. Transfer learning in minutes. See [A221] of the Performance Index for 4th Gen Intel Xeon Scalable processors. Results may vary.
7. Machine learning comparison with AMD EPYC. See [A202] of the Performance Index for 4th Gen Intel Xeon Scalable processors. Results may vary.
8. Machine learning comparison with NVIDIA A100. See [A201] of the Performance Index for 4th Gen Intel Xeon Scalable processors. Results may vary.