Don’t have my WOA book2 with me, but I think Task manager > details (6th tab) > Architecture column.
Thanks. That did the trick. As I suspected, the program was running under x86 emulation. Emulation can be glitchy at times.
Meanwhile over in MacOS on ARM land:
Wish we had more news like this for Windows. And more chips like the M2…
Not to worry…when MS/QC finally push out Nuvia v.1, Apple will only be announcing M5, the M4 successor…
And now there is this. The source is unusual given it’s a “Mac Site” but our sources on the chip manufacturing side of things corroborate that this is likely true.
It’s also supported indirectly by the fact TSMC is having significant yield issues with the 3nm node and thus two things. One, that Apple with the M3 may have TSMCs entire capacity for 3nm for the medium future, and two, that the Nuvia based Qualcomm chips are/were slated to be on a 3nm node.
I’m still a bit skeptical overall, but here you go…
PS: one of our analyst/consultants thinks this may actually be for lower performance but highly efficient chips destined for lower /mid range Surface devices such as the oft rumored WOA based Go replacement(s)
This goes to my post elsewhere that M$'s intent is use CoPilot on our silicon, the price is X; other devices (whether Windows, Android, Chromebook, Mac, iOS, iPadOS) its X+Y…
“And another aspect of Windows 12 that would benefit from in-house ARM chips would be the variety of AI features that it’s working on.”
Do you know how fab capacity is typically “rationed” in the event TSMC cannot meet its customer contracts?
In addition to Qualcomm, there is AMD who is also a major customer. Daniel Nenni guestimated Apple accounts for 25% of TSMC’s gross revenue. That’s high, but I can’t imagine that TSMC would jeopardize the rest of its customer relations just to satisfy Apple.
Jaimie Rivera had an interesting take on the M$ job postings at the end of April for ARM silicon development team - those sound like very high level leadership positions, and it only took Apple TEN YEARS to arrive at the M1, so any chance of M$ silicon before 2030?
With Apple selling 150,000,000 iPhones in 2022, not including Mac and iPad Pro sales, I don’t think TSMC would have any trepidation about hitching its wagon to Apple…
With the caveats that I don’t have a direct source within TSMC and that what information I do have comes from “background” sources, actually a chip company selling 100% of production for some period of time of a new node size and/or major architecture change is both somewhat common and not a new thing.
Apple of course was behind the most notable event when they switched to Intel from Power PC where the intel chips were both based on new architecture and were exclusive to Apple for a few months.
Qualcomm has done similar with the Snapdragon and iterations of the S series phones starting all the way back to Galaxy S4 where again that Snapdragon based on new node size was exclusive to it for about 3 months.
This also potentially benefits both vendor and customer as they can continuously optimize and or revise production during these relatively brief “exclusive” period again as the case was with the intel core chips used in the first MacBook which had a couple of significant bugs which Apple was able to "wall off " via the OS but would have been a bigger issue under Windows, but which was subsequently fixed in later revs that were made available to the other OEMs.
Thus given the ever increasing technical challenges with ever smaller node sizes and thus chip density short term “exclusives” make sense.
And again using the Qualcomm Snapdragon S4 example, I’ve heard that the initial yield rates were as low as 55%. In other words only slightly more than half of the first run of the chips actually made it in to phones.
That all being said the world and the chip market realities are very different in 2023 so this is to a degree informed speculation versus hard fact…
So what you’re saying is, it was all Apple’s from the start.
Fantastic, so in addition to software walled gardens, hardware (silicon substrate) is now entering the age of ‘balkanization’.
(Is there any way the world does not evolve into privatized domains with backroom deals governing everything…please somebody make the case.)
I sort of get the complaints, but OTOH I also see the business sense of these types of things and it would be naive to believe that for profit enterprises wouldn’t exploit every legal way to seize advantage and the shareholders (rightfully IMHO) should actually expect them to.
What does concern me a bit is that it would tend to be an impediment for the smaller participants as both Samsung and Apple are certainly able to manage the short term financial hits but a startup like Nuvia before they were acquired by Qualcomm may not have actually bring their tech to market alone.
BTW the roots of this type of behavior in the tech world go a long ways back and in fact became part of the basis of tech based antitrust legislation we do have.
eg. when General Electric and Westinghouse first were battling each other in the electric light bulb market, both sides made multiple attempts to buy the entirety of tungsten production.
I want ALL your limestone…
The site has been particularly funny today.
Semi-related. Not sure how accurate this tweet is, but it is mildly interesting.
Putting all your eggs in one basket is never a good idea.
All it could take is some major incident in Apple’s supply chain and suddenly they might not want so many SoCs.
But who is more likely to hiccup, Timmy and Apple, or Satya and Microsoft (the drop devices and fire the employee experts)?
So much for the rumor that M$ is custom building a chip to compete with Apple on ARM (or so says Daniel Rubino, who has a LOT of inside sources):