Intel’s manufacturing has crashed and the new Rocket Lake CPUs prove it
Intel Upcoming Rocket Lake CPUs Prove that it is in a crisis of existential proportions. The next six months will determine its fate and the future direction of the computer as a whole. Hyperbolic? a lot? actually no.
It’s a somewhat speculative explanation, to be sure. But Tales of the woes of Intel It’s become so routine lately that Rocket Lake’s very existence and what it means has been largely ignored. Rocket Lake says very bad things about the viability of Intel’s entire business model. This, in turn, makes it very important for the computer as a whole.
Rocket Lake is, of course, Intel’s next desktop CPU architecture. It is basically a 14nm rear port for the core 10nm Intel Sunny Cove processor architecture, as shown in Tenth Ice Lake laptop chips. Rocket Lake won’t be released until next year, which means Intel will launch a new CPU design in 2021 on the old 14nm node. Intel’s original plan was to go to 10nm in 2016. Yes, really.
Intel has sold everything “backporting” as a positive and a kind of cool, holistic approach to CPU manufacturing. “Hey guys, relax. We’re flexible, we can go from node to node. It’s free. Intel seems to say.”
The truth is, there is no such thing as a node-neutral CPU architecture. It would cost a large amount of money to bring Sunny Cove cores, PCI Express 4.0 I / O and Xe graphics to 14nm for Rocket Lake.
Therefore, it is neither awesome nor flexible. It’s a desperate move because Intel’s 10nm output node is still not good enough for prime time. Let’s repeat that. Rocket Lake will launch in 2021 at 14nm because Intel’s 10nm is still not enough to launch a desktop CPU.
This will come at a cost. As for Rocket Lake, Intel is back from 10 cores, as is evident at the moment Core i9-10900K, Goes back to eight cores. Because these Sunny Cove cores were not assigned to 14nm. They are large, creamy, and crave energy when moved down to 14 nanometers.
There are other factors that make Rocket Lake look weird. In September, Intel officially confirmed the release of the eight-core “H” of its latest 10nm Tiger Lake laptop chips Existing. It’s a CPU that makes Rocket Lake completely redundant. So why isn’t Intel releasing this chip as Desktop in place of Rocket Lake? The only plausible reason is that 10 nm is still essentially broken.
As a short-term stopping point, Rocket Lake likely makes sense, even if Intel’s marketing proposition to downgrade eight cores is almost certain to make your ears bleed. But beyond that, this “backporting” shizzle is definitely not a visitor.
Currently, Intel says its first 10nm desktop CPU will be Alder Lake, and it is slated to be introduced in the second half of next year. This is the one with the new Great hybrid engineering And up to eight performance cores and eight efficiency cores. If the Ice Lake move to 14nm came with concessions, then moving the Alder Lake to 14nm would definitely be less attractive.
In short, Lake Alder would have to be within 10 nanometers to be viable. But there are no indications whatsoever that Intel’s 10nm technology will be good enough in just over six months. It is worth noting that until now, Intel is still only selling quad-core laptop CPUs at 10nm. The launch of the Ice Lake-SP 10nm server chip, which was due in 2019, has been postponed back in early 2021. And you’d be brave assuming it hits that deadline.
What’s more, earlier this summer, Intel recognized the 7nm node, which was once promoted as the solution to all of its 10nm problems, She was a year behind schedule in what he described. 7nm won’t be operational until late 2022 at least. Being realistic, 2023 is the earliest time you will see Intel 7nm processors. And if you had to bet, you probably wouldn’t imagine 2023 much.
Meanwhile, the Taiwanese chip foundry that manages many of Intel’s competitors, TSMC, appears to be going from strength to strength. Admittedly, direct comparisons of production contract are difficult. Most observers agree that a 10nm Intel node equals 7nm TSMC for transistor density. But you can, today, buy actual consumer charging products powered by fairly large and complex chipsets based not only on TSMC 7nm but TSMC 5nm.
It’s fair to say, then, that TSMC 5nm looks much healthier than Intel 10nm at the moment. This puts TSMC not only in front but two whole nodes. This is a shocking indictment of Intel’s predicament.
In any case, the long and short term is that Intel is rapidly approaching a crisis point of existential proportions. You will likely escape Rocket Lake. But if you can’t bend 10 nanometers into some shape in time for Alder Lake in the second half of 2021, then the entire future roadmap will become unworkable and Intel will definitely have to think seriously about what was previously unimaginable. Namely, giving up making their own chips and planting them for a third-party foundry. At this point Intel is no longer what we knew it to be.
In fact, there’s a good chance Intel either has already made that call or is in the midst of overcoming it. It is possible that Intel has already decided to undertake this fundamental strategic shift, not investing the billions required to achieve 7 nanometers, but to spend that money transforming itself into an IP-based business like AMD rather than what it is today – a manufacturer who needs designs Inner slats to save those billions of dollars.
Of course, speculation about future events is often a cup game. It could take years for events to unfold. But it is different this time. Intel is fast approaching this existential crisis point. Six months from now, we’ll know if Intel has changed things. Or if the basic PC-powered tech landscape we love is about to undergo a drastic change.