Intel’s Oft-Delayed “Sapphire Rapids” Xeon Processors Finally Arrive in Early 2023

Intel’s next-generation Xeon processors based on the new Sapphire Rapids architecture have been delayed more than once, but Intel is finally gearing up to start selling them to PC companies and end users. According an-ad-on-an-ad Intel’s tweet that a “data center launch event” on January 10 will feature Sapphire Rapids processors, and the chips have currently “met product release qualifications and the company is beginning to ramp up deployment.”

Also referred to as “Xeon Scalable” or 4th Gen Xeon, Sapphire Rapids processors were originally scheduled for release in late 2021, but by mid-2021 it had become Q1 2022, then “later in the year than expected at the origin”. and now early 2023. We still don’t know when the chips will actually start shipping, only that we’ll get more information in January. These types of delays are relatively common for Intel, which has also struggled to release its dedicated Arc desktop GPUs on time and suffered repeated manufacturing setbacks over the past decade.

In an interview with The Verge last month, Intel CEO Pat Gelsinger spoke about the Sapphire Rapids delays, tacitly blaming them on previous leadership and indicating that future products wouldn’t have the same bumpy rollout.

“This project was started five years ago, so it’s ongoing. I can’t just reset the methodology for a product that started five years ago,” Gelsinger said. “[Sapphire Rapids] was far too complex, with three major new systems, or interfaces, in this design…and there was no backup on any of them.”

Sapphire Rapids is a distant relative of the Core Alder Lake (12th Gen) processors that have been shipping in laptops and desktops for about a year now, built using “Golden Cove” based processor cores (unlike in Alder Lake, Sapphire Rapids uses no low-power cores) and the same Intel 7 manufacturing process. Up to 60 cores can be expected for data center versions of the processor and up to 56 cores for workstation versions, at TDPs of up to 350W. But Intel’s delays have made the chips less competitive than they would have been had they launched earlier this year. AMD’s next-gen Epyc processors for servers (codenamed Genoa) will offer up to 96 Zen 4 cores per processor when they launch later this month, while current-gen Threadripper processors already exceed 64 cores .

Other Sapphire Rapids features include DDR5 memory support, PCI Express 5.0 connectivity, and Compute Express Link (CXL) 1.1 standard support, all features that will also be supported by AMD Genoa. (These were the “three major new systems” Gelsinger referred to in his interview.)

Architecturally, one of the most remarkable things about the chip is that it’s Intel’s first foray into chip-based processors – each processor actually consists of multiple dies in silicon, interconnected by a high-speed interconnect. AMD has used a chip-based approach for all of its Ryzen, Threadripper, and Epyc processors, and this may be a way to improve manufacturing yields; if there is a fatal flaw in a chip chip, you have to throw away far less silicon than for a similar flaw in a huge monolithic CPU chip. It also allows for mixing and matching manufacturing processes, so you can use a state-of-the-art process for the things that will benefit the most (CPU and GPU cores, for example) while using a cheaper and more efficient process. mature for other things (I/O and other chipset functions).

Intel will rely even more on chips starting with its 14th generation “Meteor Lake” processors, which will combine a mix of “tiles” built using different manufacturing processes into a single processor.