Intel Corporation Presents at COMPUTEX Taipei 2024, Jun-04-2024 11:00 AM - SEHK:4335

SEHK:4335

During the presentation at COMPUTEX Taipei 2024, Diana Wofk provided a demonstration of an AI appliance that combines Xeon and Gaudi 2 technology, utilizing components from the Open Platform for Enterprise AI (OPEA), such as a vector database. Here is a detailed list of Diana Wofk's statements:

1. Introduction to the AI Appliance Demonstration:

   • "All right. So I thought we'd do a quick demo of that AI appliance that you mentioned. So this is Xeon and Gaudi 2 with components from OPEA such as the vector database."

2. Explanation of the RAG Plus LLM Example:

   • "And we have here an example that's RAG plus LLM. And I should mention that the LMM we're using is based on LLaVA, which is a multimodal LLM. So that will come into play, which we use to the demo."

3. Starting with a Medical Query:

   • "So let me start with maybe a quick medical query."

4. Explanation of Multimodal LLM Response:

   • "So you can see like a typical LLM we're getting the text answer here, standard, but it's a multimodal LLM. So we also have this great visual here of the chest X-ray."

5. Further Analysis of the Chest X-ray:

   • "I'm not great either. But the nice thing about -- and I'm going to spare you my typing skills. I'm going to do lot cut and pasting here. The nice thing about this multimodal LLM is we can actually ask a question to further illustrate what's going on here. So this LLM is actually going to analyze this image and tell us a little bit more about this hazy opacity, such as this."

6. Additional Query on Gaudi Pricing:

   • "Okay. Well, I mean RAGs, of course, as you said, are great for up-to-the-minute information, also confidential on-prem data, so -- and I know this RAG is super up to date. So I thought maybe I would ask it a question I've been wondering about recently."
   • "Okay. So it's actually, it has to do with the prices of the Gaudi 2 and Gaudi 3 kits. So I figured I just asked that to our RAGs."
   • "So I thought I'd just ask -- what?"
   • "It might be now."

These statements highlight Diana Wofk's demonstration of the AI appliance, showcasing the capabilities of the RAG plus LLM setup and the multimodal nature of the LLaVA model used in the demonstration. She also queried the RAG system about the pricing of Gaudi 2 and Gaudi 3 kits, inadvertently revealing confidential information during the demonstration.

Patrick P. Gelsinger, the CEO and Director of Intel Corporation, presented at COMPUTEX Taipei 2024. Below is a detailed summary of his remarks:

Opening Remarks

• Expresses gratitude for the opportunity to present at Computex and acknowledges the central role Taiwan plays in the tech ecosystem.
• Highlights the significance of the event as it focuses on future innovations rather than current ones.

Moore's Law and Innovation

• Mentions the first Computex event in 1981 and Intel's introduction of the 80286 chip, which had 100,000 transistors.
• Notes the progress from the 80286 to the possibility of having 1 trillion transistors on a single package by the end of the decade.
• Affirms that Moore's Law is alive and well and emphasizes Intel's commitment to process technology leadership.

Intel's Presence in Taiwan

• Celebrates the 40th anniversary of Intel's operations in Taiwan and announces plans for a celebration at the next Computex event.

Partnership Acknowledgment

• Expresses gratitude to Intel's partners and acknowledges the collaborative efforts that are driving innovation.

AI Era and Semiconductor Industry

• Compares the AI era to the advent of the internet, predicting that every device and company will become AI-driven.
• Anticipates the semiconductor industry reaching $1 trillion by the end of the decade, driven by AI advancements.

Intel Foundry Services

• Introduces Intel Foundry Services, positioning Intel as the world's first system foundry for the AI era.
• Discusses the importance of flexible and resilient supply chains for the global economy.

Intel Products Portfolio

• Focuses on Intel's products portfolio and its role in enabling AI everywhere.
• Describes the AI era as the most consequential period of his career and highlights the pervasive nature of AI in daily life.

Data Center Innovations

• Highlights Intel's leadership in the data center with 130 million Xeons powering data centers worldwide.
• Announces the launch of Xeon 6 with E-cores, emphasizing its high-core count, high-density, and exceptional performance per watt.
• Demonstrates the efficiency and performance improvements of Xeon 6 compared to previous generations, using a live demo with Chuck.
• Discusses the sustainability benefits of Xeon 6, including significant energy savings and carbon footprint reduction.

Xeon 6 with E-cores

• Details the performance and power efficiency improvements of Xeon 6, including a 4.2x performance gain in media transcoding and a 2.6x performance per watt improvement.
• Mentions plans for a second-generation Xeon 6 with 288 cores, offering a 6:1 consolidation ratio.

Customer Testimonials

• Shares positive feedback from customers like eBay and SAP regarding performance and energy efficiency improvements with Xeon 6.

AI Workloads and Solutions

• Discusses the evolution of AI workloads, focusing on real-time databases (RAG) and their integration with large language models (LLMs).
• Highlights the complementarity of Xeon and AI accelerators like Gaudi in handling AI workloads.
• Announces the price of Gaudi 2 and Gaudi 3 AI kits, noting their compelling cost-performance ratio.

Gaudi 3

• Introduces Gaudi 3, emphasizing its performance, cost-effectiveness, and scalability.
• Showcases the "wall of Gaudi 3," highlighting the ecosystem momentum and mass market opportunities.
• Partners with companies like Inventec to deliver AI compute capabilities across various enterprise sectors.

Networking Innovations

• Stresses the importance of open networking standards and introduces AI-optimized Ethernet-based networking solutions.
• Announces the Ultra Accelerator Link for scale-up AI communication in the data center.

Edge Computing

• Highlights Intel's strong foundation in edge computing, with 90,000 edge deployments and 200 million CPUs over the past decade.
• Launches the Intel Tiber Edge software platform to simplify AI application deployment at the edge.
• Provides examples of healthcare applications, such as Samsung Medison's ultrasound solution, enabled by Core Ultra and OpenVINO.

PC Market

• Describes the current moment as the most exciting in 25 years for the PC market, drawing parallels to the introduction of WiFi.
• Predicts that 80% of all PCs will be AI PCs by 2028, with Intel leading the way.
• Announces the launch of Lunar Lake, a revolutionary design for the next generation of AI PCs, with over 80 designs and 20 OEMs.
• Discusses the CPU, GPU, and NPU capabilities of Lunar Lake, emphasizing performance, power efficiency, and compatibility.
• Compares Lunar Lake favorably against competing solutions, including X Elite, in terms of CPU, GPU, and AI performance.

Partnership with OEMs

• Invites partners like Asus and Acer to discuss their experiences with Core Ultra and Lunar Lake, highlighting customer enthusiasm and innovative use cases.
• Emphasizes the transformative potential of AI PCs and the shift from search to ask, as well as the emergence of new usage models.

Closing Remarks

• Reiterates Intel's commitment to driving innovation and enabling a digital world powered by silicon.

• Affirms belief in Moore's Law and Intel's dedication to process technology leadership.

• Thanks the Taiwan ecosystem and partners for their contributions and looks forward to continued collaboration.