Apple M1 Pro and Max Surprises

Apple M1 Pro & Max Surprises

by Jean-Louis Gassée

The rumored M1X processors that were supposed to power 14”and 16” MacBook Pro laptops turned out to be muscular M1 Max and M1 Pro chips. Intel fans reacted predictably but a more reasoned, authoritative evaluation came from an unexpected source.

The first thing to say about the M1 Pro and M1 Max, Apple’s new homegrown processors, is that we misunderestimated them. The chips were a well-guarded secret; pre-announcement scuttlebutt mentioned an “M1X” chip, christened in the manner of the A12X that powers the third generation 2018 iPad Pro.

Just as the A12X was a brawny derivative of the A12 found in the 2018 iPhone X series, with an additional core, a beefier GPU (Graphics Processing Unit), dedicated neural network hardware, and 70% more transistors, so, too, the M1X will add a little bit of everything to the well-regarded M1 SoC (System On a Chip). We’ll have a CPU ready to support next-generation MacBook Pro laptops, including a bigger-screen machine for “real” pros who are writing code or developing high-bandwidth media content.

That was how the rumors ran, anyway. But the reality that was unveiled in Apple’s “Unleashed” event last Monday was filled with surprises.

The initial surprise was that the silicon birth gave us twins: two new M1 CPUs. Furthermore, the M1 Pro and M1 Max aren’t just “a little bit more of everything” derivatives of the original M1, they’re impressively muscular engines. Regard the transistor count: 33.7 billion and 57 billion respectively for the M1 Pro and M1 Max versus a “mere”16 billion for the original M1. An authoritative, well-worth-the-read AnandTech article provides a graphic that helpfully compares the three devices:

The Intel side of our village has dismissed the M1 Pro and Max as impressive but hardly threatening: “Sure, Apple has a fleeting advantage due to their access to TSMC’s denser 5 nanometer process, but once Intel gets there, x86 chips will outperform Apple Silicon, especially with their access to the vast library of Windows software.”

Some things never change. Intel fans had the same reaction, eight years ago, when Apple introduced its first 64-bit processor, the A7 that powered the iPhone 5. Reminiscing about age-old IBM dezinformatsiya, I had fun with the competitor’s pouts in a 2013 Monday Note titled 64 Bits. It’s Nothing, You don’t need it. And We’ll Have It In 6 Months.

An extreme case of Just You Wait! rhetoric can be found in a Seeking Alpha piece titled Apple M1 Max: Less Impressive Than It Looks. The author, who has a vested interest in Intel stock, summarizes his points thus [as always, edits and emphasis mine]:

“• Apple has launched new M1 Pro and Max chips. These are basically scaled-up M1s (which was a scaled-up A14).

• Even based on Apple’s own performance estimates, the chip will be slower than Intel’s upcoming Alder Lake CPU. Neither is the M1 Max faster than Nvidia in GPU performance.

• Hence, Apple’s main achievement remains its power efficiency instead of performance. However, this is (at least partly) attributable to the chip’s process advantage credited to TSMC.”

The faithful author then falls for a neat Intel markitecture trick:

“As Intel said, the SiP is the new SoC, so from that view, a large 400mm2 SoC should be seen as a point of weakness (large chip = large manufacturing cost) rather than strength.”

In the case of x86 devices, Intel’s SIP (System In a Package) is an admission of their inability to integrate all of the CPU organs into a single SoC. As a result, SIP performance suffers because of its lower interconnect speeds, particularly when compared to a fully integrated SoC. For example, the memory transfer offered by the M1 Pro and Max reaches 200 and 400 gigabits per second, speeds that are unattainable on a SIP CPU implementation.

For a sober and objective counterpoint from an unexpected source, read Steven Sinofsky’s extensive Twitter thread which he obligingly summarizes in a Medium article titled Apple’s Long Journey to the M1 Pro Chip. In tweet 25 from the thread, Sinofsky summarizes the advantages of Apple’s new chips:

“25/ When you look at M1 Pro/Max today it is tempting to think of this in terms of performance, but performance per watt AND integrated graphics AND integrated memory AND integrated application processors is innovation in an entirely different direction. Just the beginning.”

Steven Sinofsky’s background gives a special resonance to his Apple processor commentary. Before retiring from Microsoft after more than two decades at the company, he was Windows President. In that position, Sinofsky was intimately aware of the ins-and outs of Intel’s x86 development and operating systems support challenges. In other words, a unique vantage point from which to evaluate Apple Silicon.

(Sinofsky is the author of Hardcore Software Inside the Rise and Fall of the PC Revolution, an immense online opus compiled with Benedictine patience and extensive archives. A great read for those who have lived inside the PC revolution, or want to gain a better appreciation for its unfolding.)

Another surprise at the “Unleashed” event was the announcement of the “containers” for the new CPUs. This is no place for a recitation of the full feature set for the new MacBook 14- and 16-inch Pros, but a few important changes deserve attention, changes that some see as an admission of fault, a sort of Jimmy Swaggart I Have Sinned Against You confession.

Gone is the much-maligned Touch Bar, replaced by real function keys and, mirabile visu! we now have three Thunderbolt 4 (USB-C) instead of two, an SDXC card reader, an HDMI port and…the return of the Magsafe charging connector! (Charging is also possible through a USB-C connector.)

All of the features that the Apple schoolmarm kept saying we didn’t need have now returned, and then some. Although the usual naysayers see unwanted Product Manager fingerprints all over their formerly pristine machines, most reviewers and patrons (myself included) are delighted and relieved by the “retro” moves. In particular, the new connectors are a welcome change from the company’s previous hauteur. Goodbye Dongletown!

When Apple announced the M1 processor at the June 2020 WWDC, it was seen as a straight-line evolution of the Ax processors that powered iPhones and iPads. For several iterations, the Ax had shown “desktop-class” computing power, so the new M1 felt like a logical replacement for the x86 on a MacBook Air or a 13” MacBook Pro.

Jumping this evolution forward to its furthest stage, inquiring minds wondered if Apple would someday be able to replace the monster Xeon Intel chip that powers the Mac Pro…

I think we now have the answer.

In some no-too-distant future an M2 chip will complete the two-year transition announced in June 2020. Let’s just hope a new M2 Mac Pro has a starting price that’s less than the current $5,999 — with extras that cost as much as $7,000 for a 28 core Xeon W processor, $11,660 for Radeon Pro W6900X GPUs, to say nothing of the $299 Feet or $699 Wheel Kits. I’ll stop before getting into $5,999 Pro Display XDR territory and its famous $999 stand…

Seriously, completing the transition away from Intel chips will raise many interesting questions, including Microsoft’s participation in the transition and the impact on PC OEMs.

I’ll leave the last words to Steven Sinofsky: Just the beginning.

— JLG@mondaynote.com

from Hacker News https://ift.tt/3jB4mxo