Intel’s Server Chips Are ‘Lead Vehicles’ for Manufacturing Strategy

By Agam Shah

March 30, 2023

…But chipmaker still does not have an integrated product strategy, which puts the company behind AMD and Nvidia

Intel finally has a full complement of server and PC chips it will release in the coming years, which will determine whether it has regained its leadership in chip manufacturing.

The chipmaker this week announced a new processor called Clearwater Forest which will be the first server chip made on its most advanced process called 18A. The process has next-generation transistors and other technologies that will make chips faster and more power-efficient than predecessors.

Intel 18A is the process the chipmaker hopes will give it a manufacturing advantage over rival Taiwan Semiconductor Manufacturing Co., which boasts AMD, Apple, and Nvidia as its clients.

For that, Intel has taken on the challenging job of advancing five nodes in four years, which would have otherwise taken eight to ten years. Intel has invested billions in new factories to achieve that goal.

What’s Coming Up

Clearwater Forest will come to market in 2025, and will be “the culmination of our five nodes in four years strategy,” said Sandra Rivera, executive vice president and general manager for the datacenter and AI group at Intel, during a presentation of Intel’s roadmap this week.

The Clearwater Forest chip will have low-power Xeon “E-Cores,” and are designed for hyperscalers, which typically need higher core counts to serve cloud-native applications. It will succeed a server chip called Sierra Forest, which is the first generation of low-power Xeon cores for hyperscalers.

Intel’s Xeon chips are what Rivera termed “lead vehicles” for most of the upcoming nodes.

The high-margin server chips will serve two purposes as lead vehicles: it will prove Intel can make high-performing chips on new nodes, which should ease customers that have doubts about the chipmaker’s ability to deliver products on time and reach its manufacturing goals, said Jim McGregor, principal analyst at Tirias Research.

With final products announced for each node, Intel now has no room for error, and one misstep can affect its credibility and have the domino effect for its server roadmap to come crashing down, McGregor said.

Later this year, Emerald Rapids will succeed the current Sapphire Rapids and will be released on Intel 7. Intel’s PC chip Meteor Lake will be released on Intel’s next node called Intel 4, which will be the first to use EUV technology to etch finer features on chips.

That will be followed by server chips Sierra Forest and Granite Rapids on Intel 3, which will be released next year. A PC chip called Arrow Lake – due for release next year, will lead off Intel’s 20A, and Clearwater Forest will lead off the most advanced 18A node. Intel has also listed a PC chip called Lunar Lake, which will likely be made using 18A.

Source: Intel Corp.

Server Chips Are Priority

In the past, Intel’s release of the first chips on new nodes has swung between Xeon chips and PC chips, and that matters because it reflects high-priority markets for the chipmaker.

“Even when PCs go through highs and lows, servers are still ramping because of new applications. You do not want to miss out on hyperscalers, AI, and other markets,” McGregor said.

Intel’s Rivera said that server chips provide a 40% to 50% margin, and the cash from server chip sales will help alleviate the costs related to building the factories, to which Intel has committed billions of dollars.

“We have to pay for all of that process technology and you’ll see that in terms of the operating margin,” Rivera said.

The disaggregated chip design based on tiles, also called chiplets, changes the mindset of how to construct processors, as it allows dies from older nodes to be combined with cutting-edge server CPUs, Rivera said.

“Not every block – memory block, I/O block, different parts – of the IP infrastructure may benefit from the latest process nodes. We are going to be smart about how we design the products, how we architect them, how we bring them to market, lowering the risks, lowering the costs, and increasing the predictability and quality of our execution,” Rivera said.

Timely releases of Intel’s server chips on new nodes may convince external customers that the factories are running at full steam and ready to make custom chips.

Intel is targeting hyperscalars that will buy customized chips in large volumes, and similarly, previous attempts by the chipmaker to open up its factories to external customers were only targeted at large customers, McGregor said.

Intel’s Non-Integrated Chip Strategy

The server roadmap presentation also outlined a disaggregated approach in product development, with each product unit developing products independently, with few tie-ins. Rivera said more than 15 new FPGA products are scheduled to appear this year, and Gaudi3 is in the pipeline.

That is unlike rivals like AMD and Nvidia, which are building comprehensive product lines that tie together CPUs, GPUs, FPGAs, AI chips, and data processing units.

AMD and Nvidia are seeing success with integrated offerings, but Intel’s future is in manufacturing custom chips for clients, so it makes sense to keep IP blocks separate rather than integrate them, McGregor said.

Intel is learning from the experience with Sapphire Rapids, which it spent three years building, with many changes to meet the fast-changing AI needs.

“Intel has done that with Sapphire Rapids. It is a CPU with accelerators. It is going to have so many different SKUs with memory and accelerator combos,” McGregor said, adding “it helps to have independent IP and if one gets delayed, you aren’t held up.”

The changes to Sapphire Rapids took many years, which is what delayed the chip’s release. Such changes will be easier with chiplet design, which brings in the ability to swap out processing tiles, and a flexible design helps mix and match CPUs with internal or external accelerators from business unit.

Source: Intel Corp.

Rivera gave the example of the AI market, in which inferencing is evolving faster than training as generative AI models are deployed. Intel has three chip options, which include the Xeon Max GPU, Gaudi chips, and Agilex FPGAs. The product offerings are independent of each other, and Rivera is following the strategy established by CEO Pat Gelsinger: to chop paperweight products and refocus on engineering.

Intel has already simplified its high-performance computing roadmap by canceling the Rialto Bridge GPU, which was the intended successor to the Ponte Vecchio chip.

Intel also last month changed plans for its chip codenamed Falcon Shores, which was to combine an x86 CPU and Xe GPU into a chiplet architecture. Instead, Intel will first release a GPU-only version of Falcon Shores and will do so in 2025 rather than the originally planned debut in 2024. Intel has talked to customers and is revamping its roadmap accordingly, a company spokeswoman said.

“Running a playbook where all teams are focused on our commitments prioritizing critical aspects of our programs and executing at a predictable cadence will enable us to be more competitive in high-growth parts of the market,” Rivera said.

Fixing a Broken AI Strategy

The disaggregated server chip strategy is also an effort to undo Intel’s haphazard approach to AI, which is built around the dozen or so companies it has acquired, McGregor said.

The company acquired chipmakers like Nervana, which failed, and is now building its AI training strategy around Habana, which makes the Gaudi chip. The Xeon Max GPU, which is targeted at high-performance computing, will be an option for AI and deep learning. The Agilex FPGAs and eASICs are targeted at inferencing.

But Intel needs to figure out how it wants to sell its product, and there is a good chance that a bouquet of independent chips may not work.

“Everyone’s trying to rearchitect the datacenter and trying to upgrade the network with the changes in what we’re seeing, especially around AI. It’s changing so rapidly, it is kind of like Intel needs to have an integrated strategy,” McGregor said.

Intel’s market share in server chips was 82.4% in last year’s fourth quarter, a decline from 89.3% in the same quarter of 2021, according to Mercury Research. AMD’s server market share was 17.6% in the fourth quarter last year, growing from just 10.7% in the fourth quarter in 2021.

Header image: Lisa Spelman, Intel corporate vice president and general manager of Intel Xeon Products, displays a wafer holding Sierra Forest processors. (Credit: Intel Corp.)

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion XL — were added to the benchmark suite as MLPerf continues Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire