Intel GPU roadmap reveals hybrid integrated discrete graphics card • The Register

Intel on Thursday presented investors with its future roadmap for graphics accelerators, which plans to blur the lines between integrated and discrete graphics.

The chip giant said it will ship its discrete desktop Arc GPUs, codenamed Alchemist, in the second quarter of this year. Arc GPUs will appear in laptops by then in Q1. Workstation-grade components are expected in Q3.

“We have received a very enthusiastic response from our partners who have already won more than 50 designs from major OEMs and add-in board vendors,” said Raja Koduri, senior vice president and general manager of Accelerated Group. Computing Systems and Graphics at Intel, in a recorded speech for the Investor Day Conference.

Alchemist discrete GPUs will ship with Alder Lake and Raptor Lake processors to be unveiled by Intel this year. Intel’s Arc GPUs will get really interesting starting next year with its GPU codenamed Battlemage, which will ship with the upcoming Meteor Lake processor and its next-generation CPU architecture.

Battlemage GPUs will be integrated as an array of chips — Intel calls them tiles — alongside other chips containing processor cores and supporting circuitry, all contained within a single large Meteor Lake processor package. This approach comes as Intel moves to advanced manufacturing nodes in which compute tiles can be stacked vertically and linked by high-speed interconnects.

“Meteor Lake is a brand new architecture that will allow tiled GPUs to be integrated on a 3D package. This is super exciting, as it allows us to deliver discrete graphics-class performance with the efficiency of integrated graphics,” Koduri said. It’s a new class in graphics that can’t be called integrated or discrete, he said.

“This is just the beginning of the strategic advantages that this tiled architecture will give us. We’ll talk more about that in the future,” Koduri said.

Intel may have an early advantage over tiled GPUs as it will be the designer and manufacturer of its own chips. Nvidia is a fabless company that relies on third-party fabs and hasn’t talked about developing tiles to slot alongside x86 cores in future PC chips.

Nvidia did not respond to request for comment on its plans to develop tiles to complement its discrete and integrated GPU products.

Intel outlines the coming years

Intel has already started working on its graphics architecture for beyond 2024, code name Celestialwith the goal of achieving graphics leadership “in every segment we compete in, from low-power mobile workstations to high-performance workstations,” Koduri said.

Intel’s graphics division is continuing the metaverse with a software project called Endgame, which will provide users and developers with access to Arc GPUs as “a continuous, always-on, low-latency, persistent, and immersive computing service,” it said. said Koduri.

Intel will share more details about Endgame this year. But reading the tea leaves, Arc GPUs could be used as bare metal in PCs or remotely in the cloud to run metaverse applications that include gaming, animation, and visualization. Nvidia offers this through its Omniverse Platform, a largely closed metaverse platform that includes hardware and software products from Nv.

“With the combination of our silicon platforms with Endgame, we anticipate that Intel and the PC platform will continue to thrive as we usher in the metaverse,” Koduri explained.

The executive leads a division formed last year and is targeting the broader accelerated computing market, with processors not sufficient to meet the computing needs of graphics, supercomputing and artificial intelligence.

The division’s product roadmap includes supercomputing chips such as Ponte Vecchio. Koduri is targeting $10 billion in revenue by 2026 through its initiatives. The total addressable market opportunity will reach $155 billion by 2026, it is hoped.

“We are focused on the $100 billion opportunity in silicon platforms. Our strategy is to collaborate and openly participate with the ecosystem for the remaining $55 billion in system software and services,” said said Koduri.

It then attempted to take a shot at rival Nvidia, which dominates the supercomputing and AI space and has an early lead in the metaverse opportunity.

“This is in stark contrast to the approach taken by our main competitor today. They aim to eat away at the ecosystem. Although their closed proprietary approach may have short-term benefits. We do not believe a closed approach is scalable long term running for this big opportunity,” Koduri said.

Nvidia did not respond to requests for comment on Koduri’s quote.

Intel’s visual computing roadmap… how and when to expect next-gen GPUs. Click to enlarge

In a separate speech, Intel’s senior vice president of software group Greg Lavender said the company is the largest contributor to the Linux kernel and employs more than 120 maintainers of key open source technologies. Lavender also reaffirmed Intel’s commitment to open source with things like its OneAPI.

But in recent years, the company has scaled back its Open Source Technology Center, its public face of the open source movement. The register has reported extensively on Intel’s secret Software Defined Silicon (SDSi) initiative in which Intel staff members have posted hints on the Linux kernel mailing list regarding new features in which users may have to purchase licenses to enable features physically present in processors not available for use out of the box.

Intel is a wonderful member of the open source community and contributes extensively to projects with OneAPI, networking components and more, said Michael Larabel, Linux software engineer and founder of Phoronix, which tracks Linux kernel contributions.

“Intel’s secrecy about enabling new features isn’t new, but more often than not their work on new features tends not to be controversial like SDSi,” Larabel said. The register.

Larabel has mixed feelings about SDSi, as Intel hasn’t provided preliminary details, which has sparked speculation and concern.

“It looks like future processors with SDSi are arbitrarily restricting certain features until after the fact ‘license’ will be pursued and what pricing will look like,” he said.

Intel tries to release kernel source code for new hardware support and feature enablement early before chip launches, which means it’s not necessarily very open with all the details. Its current goal may be to conceal these features to avoid ruining launch buzz or competitor analysis, Larabel commented.

“So while initially some areas may not be effectively communicated with early patches, at least support is working ahead of time to allow time for integration into the Linux kernel and major distributions,” a- he concluded. ®