Intel held a release party on Tuesday for its Meteor Lake, the processor that goes on sale December 14. But if you’re trying to decide whether you should stick with Intel-powered Windows laptops or upgrade to Apple’s efficient and powerful new MacBooks. , pay attention to the other three processors in preparation.
At Intel’s innovation conference, CEO Pat Gelsinger touted a series of new processors expected to arrive in 2024 and 2025. The first are Arrow Lake and Lunar Lake. Then, in 2025, will come Panther Lake, whose design is “well underway,” Gelsinger said, confirming a rumored code name.
In an effort to show that the products are not vaporware, Gelsinger showed off a prototype Lunar Lake computer, an unsightly blue box marked “Lab CSF” and featuring technical controls such as “clear cache” and “virtual battery.” » which you will not see on your computer. Average PC. And he showed a wafer built with Intel 18A, the manufacturing process that the company hopes will restore the chipmaking leadership it lost to Taiwan Semiconductor Manufacturing Co. (TSMC) and Samsung.
See also: In Intel’s chip factory, I saw the future. It’s plain old glass
These and other technology demonstrations bode well for a company that has struggled for years, according to James Sanders, an analyst at CCS Insight. “You can definitely say there’s a sense again that Intel is an engineering-led company again,” Sanders said. “That’s the image they have to give after being run for years by accountants.”
Intel has supplied Apple with processors for its Macs for years, but Apple kicked out Intel and moved to its own M series of processors, more powerful variants of the same chips that power iPhones and iPads. The MacBook M1 and M2 were praised for their speed and long battery life, and Intel executives recognized that improving both attributes was a top priority. Intel hasn’t revealed Meteor Lake performance details, but the company promises improvements in processing, graphics, and AI performance.
For the progression of processors from Meteor Lake to Panther Lake, Intel compares its performance to that of its competitors, including Apple, on three speed tests: CPU (central processing units for general computing), GPU (graphics processing units ) and NPU (neural processing units to accelerate AI). ) within a fixed power budget.
“We’re looking at the overall capacity that we offer between these three platforms, and we think these platforms are becoming very competitive, the best that Mac or anyone else offers,” Gelsinger said during a press conference at the show. “We are very, very happy with the roadmap.”
Apple did not immediately respond to a request for comment.
Intel factories will start manufacturing Panther Lake by March 2024
Intel manufacturing facilities, or fabs, will begin building the first Panther Lake processors in the first quarter of 2024, Gelsinger said.
“This is Intel 18A, the finish line of our five nodes in four years,” he said, referring to the company’s efforts to rapidly move toward new manufacturing process improvements in order to become competitive again.
Intel also showed off new Xeon processors for the giant data centers that run the tech giants’ cloud computing services, announced that Stability.ai would buy its Gaudi AI accelerators, and touted new substrate technology of glass that he says will make processors faster and more energy efficient. and larger later in the decade.
AI processors are the largest processors on the market, and demand has increased with the explosion of interest in generative AI technology. Intel’s competitor Nvidia was the biggest beneficiary, but Intel announced that Stability.ai, an AI generative imaging company, would buy a Gaudi-based AI supercomputer with Xeon processors overseeing 4,000 Gaudi’s AI accelerators.
New steps in the manufacturing of Intel chips
Intel is catching up with its competitors in one area of chipmaking, that of using extreme ultraviolet (EUV) light to etch finer details onto silicon wafers. He expects to be ahead, at least a little, with the next breakthrough, called high numerical aperture (high NA) EUV.
This should allow chipmakers to pack even smaller features and fit more transistor circuits onto a processor. But the higher level of detail with the more advanced EUV process comes at a cost: the maximum CPU size it can create is half that of a typical EUV. This is primarily a problem with large processors like AI accelerators, and processor manufacturers may bundle multiple “chiplets” into a single package to compensate.
Intel expects to get the first EUV machine for high-volume chip manufacturing by the end of 2023, Gelsinger said. The only supplier is the Dutch company ASML. Its Intel 18A process could use high NA, but 18A chips could also be made with conventional EUV and more steps in the manufacturing process.
Gelsinger has been talking about the plan for five nodes in four years for more than two years: Intel 7, Intel 4, Intel 3, Intel 20A and Intel 18A. At Innovation, he outlined future advancements in processor manufacturing for the first time, showing a roadmap with three additional steps beyond 18A.
The Intel 20A and 18A use a technology called Gate All Around, which Intel calls RibbonFET, to miniaturize transistors and help comply with Moore’s Law. They also use rear-end power delivery, which Intel calls PowerVia, to improve power management efficiency. Intel is working on new versions of both.
“We’re making new improvements to the gates all around,” Gelsinger said. “We are also well underway on the next version of PowerVia.”
Gelsinger says there are three types of chipmakers: “You’re big, you’re niche, or you’re dead.” Intel is trying to accelerate its progress in chipmaking to avoid being sidelined and relegated to making older chips, a fate that IBM and AMD have faced in recent years.
“Intel is way too big to be niche,” Gelsinger said, “so we better be really big.”