Will We Have an Apple M1 Max Duo and Quad in 2022? Intel-Based Solutions to Become Performance Dinosaurs Overnight?
In 2022, all Mac and PC Intel-based solutions might look like dinosaur fossils, in performance terms.
Today’s Apple M1 Max chip is already 1.7X faster for the same power draw compared to Intel laptop chips. Or just as fast at 70% less power. You can quibble over details, but the gap is huge.
Now assume that Apple can deliver an M1 Max Duo and M1 Max Quad design for the new iMac 5K and Mac Pro. Nothing else on the market would be remotely competitive.
And while Apple is surely not targeting the cloud computing market, one has to wonder what an M1 Max architecture could do for the server farm space, where computing power per watt drives everything. Given Apple’s huge presence on the internet and its 'green' marketing bent, it would make some sense to at least engineer its own M1 Max derived server farm setup, if only to chop power usage by 70%.
Reader Shameer M writes:
Mark Gurman (Bloomberg Tech reporters) posted an article earlier this year with respect to new Macs that Apple would be coming out in 2021 and in 2022. This is what he said about the new 14" / 16" MBP back in May:
"For the new MacBook Pros, Apple is planning two different chips, codenamed Jade C-Chop and Jade C-Die: both include eight high-performance cores and two energy-efficient cores for a total of 10, but will be offered in either 16 or 32 graphics core variations."
Jade C-Die (10 core CPU and 32 core GPU) is what we now know as M1 Max and Jade C-Chop (half of the GPU cores "chopped" off) is M1 Pro.
Here is what he goes on to say about the new Mac Pro coming out some time next year:
"Buyers of the high-end Mac Pro desktop planned for next year will likely have a choice of two processors that are either twice or four times as powerful as the new high-end MacBook Pro chip. Codenamed Jade 2C-Die and Jade 4C-Die, a redesigned Mac Pro is planned to come in 20 or 40 computing core variations, made up of 16 high-performance or 32 high-performance cores and four or eight high-efficiency cores. The chips would also include either 64 core or 128 core options for graphics."
Jade 2C-Die is essentially two M1 Max SoC's on a single die or package and Jade 4C-Die is four M1 Max SoC's on a single die or package Apple (AAPL) Readies MacBook Pro, MacBook Air Revamps With Faster Chips - Bloomberg
Here is a diagram someone made to highlight it: https://hypercritical.co/2021/05/21/images/city-of-chiplets.png
[MPG: not a realistic design, 4 chips together simply replicated identically does not work; that would be a multi-chip design on a single piece of silicon and would require a lot of extra logic to deal with synchronization issues. But it shows the scale of things]
If Gurman is right and the new Mac Pro is coming out next year, I don't think the processors that power it will be built on a 3nm node since TSMC has delayed 3nm products shipping to clients beginning in 2023: TSMC Roadmap Update: 3nm in Q1 2023, 3nm Enhanced in 2024, 2nm in 2025
MPG: should be a VERY interesting 2022 for performance in computing.
But I am skeptical that the Apple M1 Max (already a huge chip witih 57B transistors) can be made more than even 2X larger—the yields would drop in a major way. And a 4X scaling is not realistic at all. So I don't see how this M1 Max Duo/Quad thing can be realized—there is no existence proof that it can be done in any commercial quantity (yield = cost).
Roy P writes:
Makes a lot of sense. Remember, there is a heck of a lot more real estate at a board level in an iMac, not to mention far more ability to keep it cool... There are other things they can do for faster rendering on the display. They can put other silicon engines for gaming, video, image, 3D rendering, audio, etc. that may not be practical in a notebook. Heck, they could put four M1 Maxes into an iMac!
They don’t need to go to 3 nm, which, not surprisingly, is “delayed”.
They can do a lot more with the current architecture, through a combination of things they can do at both a chip and a board level.
We’re at the beginning of a new technology cycle, and there are tens if not hundreds of billions of dollars of net margin Apple can get out of this generation, with no one else able to compete with them, and there is no risk of anyone leap frogging them to 3 nm.
It would not surprise me in the least if Apple already had an initial set of designs done for an M2 family of chips, all done and being tested and debugged in simulations, working with TSMC, concurrently with their process development and optimizations. So when 3nm goes to HVM, Apple will be ready to pounce. My guess is, that would be 2023, this time, maybe even 2024. No hurry.
Give the market what they will be happy with, and they can’t get anywhere else, as it is. Why give them something far more, just because you know how to? By stretching it out, they get more customers to migrate to the M1x generation, and not bypass the M1x generation and go directly to the M2x generation.
MPG: so if and when we see 2X to 4X the performance of the M1 Max (I don't see how this can happen in 2022 but I hope to be wrong), I’m not quite sure why I’d need a faster computer for, say, 10 years. Certainly not for still photography. If it’s taking me 15 minutes to run a large Topaz Gigapixel AI scaling and that drops to 1 minute, “I’m good”. Because that’s an extreme case.
More likely, a superfast computer will have software bottlenecks that make real-world performance far less than the available computing power. This has long been the case, and it’s not going to change. But some long-running scalable tasks will see huge benefit, and those are the ones that count the most.