Intel Unveils Next Generation AI Solutions, Xeon 6 and Gaudi 3
Intel Gaudi 3 AI Accelerator
HOLIDAY NEWS: As AI continues to revolutionize industry, companies increasingly need infrastructure that is cost-effective and can enable rapid development and deployment.
To meet this demand head-on, Intel launched the Xeon 6 with Performance-cores (P-cores) and Gaudi 3 AI accelerator, reinforcing the company's commitment to providing powerful AI systems with optimal performance per watt and total cost of ownership. ownership/TCO) is lower.
“The demand for AI is resulting in a massive transformation in the data center, and the industry wants choice in hardware, software and developer tools,” said Justin Hotard, executive vice president and general manager of the Data Center and Artificial Intelligence Group, Intel.
“With the launch of Xeon 6 with P-cores and Gaudi 3 AI accelerator, Intel is delivering an open ecosystem that enables our customers to implement all their workloads with greater performance, efficiency and security.”
Introducing Intel Xeon 6 with P-cores and Gaudi 3 AI accelerator
Intel's latest advances in AI infrastructure include the presence of two major updates to its data center portfolio, namely:
Intel Xeon 6 with P-cores: Designed to handle demanding workloads with exceptional efficiency,It features a larger core count, doubles the memory bandwidth, and includes AI acceleration capabilities on each core. These processors are designed to meet the demands of AI performance, from edge environments to the data center and cloud.
Intel Gaudi 3 AI Accelerator: Specifically optimized for large-scale generative AI, Gaudi 3 is equipped with 64 Tensor processor cores (TPC) and 8 matrix multiplication engines (MME) to accelerate deep neural network computing. It includes 128 gigabytes (GB) of HBM2e memory for training and inference, and 24 200 Gigabit (Gb) Ethernet ports for scalable networking.
Gaudi 3 also offers seamless compatibility with the PyTorch framework, and advanced Hugging Face transformer and diffuser models. Recently, Intel collaborated with IBM to run the Intel Gaudi 3 AI accelerator as a service on IBM Cloud. Through this collaboration, Intel and IBM target to reduce the total cost of ownership in leveraging and scaling AI, while increasing performance.
Improving AI Systems with TCO Benefits
Running AI at scale requires considering a number of things, such as flexible application options, a competitive price-performance ratio and easily accessible AI technology.
Intel's powerful x86 infrastructure and extensive open ecosystem enable enterprises to build high-value AI systems with optimal TCO and performance per watt. Note that 73% of GPU-accelerated servers use Intel Xeon as the host CPU.3 Intel partners with leading OEMs, including Dell Technologies and Supermicro, to jointly develop systems specifically designed for specific customer needs to make AI applications more effective. Dell Technologies is currently designing a RAG-based solution utilizing Gaudi 3 and Xeon 6.
Bridging the Gap from Prototype to Production with Co-engineering
Transitioning generative AI (Gen AI) solutions from prototypes to production-ready systems has several challenges in real-time monitoring, error handling, logging, security and scalability. Intel addresses these challenges through co-engineering efforts with OEMs and partners to provide production-ready retrieval-augmented generation (RAG) solutions.
The solution, developed on the Open Platform Enterprise AI (OPEA) platform, integrates OPEA-based microservices into a scalable RAG system, optimized for Xeon and Gaudi AI systems, designed so customers can easily integrate applications from Kubernetes, EdHat OpenShift AI, and Red Hat Enterprise Linux AI.
Expanding Access to Enterprise AI Applications
The Intel Tiber portfolio offers business solutions to address challenges such as access, cost, complexity, security, efficiency and scalability across AI, cloud and edge environments. Intel Tiber Developer Cloud now provides an Intel Xeon 6 preview system for technology evaluation and testing. Additionally, select customers will also get early access to Intel Gaudi to validate deployment of AI models, with Gaudi 3 clusters starting to roll out next quarter for large-scale production runs.
New service offerings include SeekrFlow, Seekr's end-to-end AI platform for developing trusted AI applications. The latest updates include the release of the latest Intel Gaudi software and Jupyter notebooks that have been enhanced with PyTorch 2.4 and Intel oneAPI and AI tools 2024.2, which includes the latest AI acceleration capabilities and support for Xeon 6 processors.
(MMI)