Microsoft is changing the game. How will Apple and M chips fair?

[jlc1978]

This is actually pretty good for Apple too. It may enable virtualisation of Windows for consumers will a proper license (not depending on Insider builds) and, if they partner with Apple, the return of Bootcamp. In terms of software, it opens another door for developers to compile for ARM and may facilitate porting of software to/from Mac.
Interesting times ahead

True, but I doubt Bootcamp will return. Virtualization is likely the path forward; as Bootcamp was mainly a way to attract switchers, something not really needed now. Based on running the developer's ARM/Win, VMs will be good enough for most users; gamers won't find it useful but then again neither was Bootcamp for gamers that were playing cutting edge games.

 

[ADGrant]

A few nice things from Build 2022

- Devkit which basically looks like a Mac Mini powered by Snapdragon running WinOnArm
- Full Visual Studio 2022 on Arm64 (including Visual C++)
- A complete dev toolchain for Arm, including VS Code, Modern .net 6, Java, etc

Project Volterra is Microsoft's new Snapdragon-powered Windows 11 mini PC

Project Volterra could be your next great dev PC

www.techradar.com

The only interesting announcement in that list is Full Visual Studio 2022 on Arm64 (including Visual C++).

VS Code, .NET 6 and Java are already available in ARM native versions for Apple Silicon Macs. You will be able to run the full Visual Studio 2022 in a Windows VM on an Apple Silicon Mac. Something only an Intel Mac could do with earlier versions of Visual Studio.

Amusingly, Visual Studio 2022is also the first version of that IDE to fully support 64bit Intel CPUs.

 

[ian87w]

...is this something people are actually worried about? Seems like a very dumb concern.

Imo personally, it should be. Apple has a track record of being too easily distracted that their lead in their products/tech got nullified fairly quickly.
For example, the 2nd gen Macbook Air. Apple had a huge lead, but they stuck with it for so long, being distracted in trying to make a thinner Pro laptop, that literally every PC ultrabooks became lighter and had better displays. Plenty of other distractions that caused Apple to "lose" years of product cycle. Trashcan Mac Pro, Butterfly keyboard, touchbar. Even in the software. Look at the buggy iOS15 due to Apple being distracted on CSAM.

Currently, even the M1 is already 2 years old. Apple still has a lead, but it's not like everybody is resting on their laurels. And going back to services, chasing Netflix is a good way to burn money. I mean look at Netflix itself, they're bleeding. Those Original shows take a lot of money to make. Meanwhile, we see iCloud languishing and core Apple products facing huge delays. These kind of things can cause a setback for Apple financially. The supply chain issues and litigations from various countries' governments will take a lot of money from Apple. The last thing Apple needed is wasting more money into virtue signalling TV Shows.

 

[exoticSpice]

Look at the buggy iOS15 due to Apple being distracted on CSAM.

You do know Apple has multiple departments. From 2017 onwards Apple formed the Pro Workflows Team they worked on products such as the iMac Pro, Mac Pro, MacBook Pro 16" 2019, MacBook Pro 14"/16" 2021 and lastly the Mac Studio.

All these Mac's followed one philosophy to put function first and form second. It was ports and more ports and great thermals design with these Macs.


CSAM has nothing do with iOS 15 being buggy, Android 12 was bugs galore as well more so than iOS 15. The main effect was working from home.

Trashcan Mac Pro, Butterfly keyboard, touchbar.

The people who made these and managed these have left Apple. In fact the creator of the butterfly keyboard tried to steal Apple IP.

Currently, even the M1 is already 2 years old. Apple still has a lead, but it's not like everybody is resting on their laurels. And going back to services, chasing Netflix is a good way to burn money. I mean look at Netflix itself, they're bleeding. Those Original shows take a lot of money to make.

Do you really think the people working on Apple Sillcon work on making TV shows and movies for Apple TV+?

Heck no. They are seperate departments. The Apple Sillcon team will NEVER step foot into the Apple TV studios.

 

[exoticSpice]

Meanwhile, we see iCloud languishing and core Apple products facing huge delays. These kind of things can cause a setback for Apple financially. The supply chain issues and litigations from various countries' governments will take a lot of money from Apple.

Apple is not the only company facing supply chain issues and huge delays. The whole industry is and will face for the next few years.

Look at AMD I can't even buy a 6000 U series laptop from AMD. Microsoft also focuses on multiple vuntures its services are not movies but games. MS spends huge amounts on xbox and also spends huge money to buy studios.
Recently MS spent $60 billion on one single game studio.

 

[exoticSpice]

The last thing Apple needed is wasting more money into virtue signalling TV Shows.

Apple TV shows are good. Have you seen Ted lasso and the latest dinosaur doc. I didn't find virtue signalling there.

 

[thenewperson]

Imo personally, it should be. Apple has a track record of being too easily distracted that their lead in their products/tech got nullified fairly quickly.
For example, the 2nd gen Macbook Air. Apple had a huge lead, but they stuck with it for so long, being distracted in trying to make a thinner Pro laptop, that literally every PC ultrabooks became lighter and had better displays. Plenty of other distractions that caused Apple to "lose" years of product cycle. Trashcan Mac Pro, Butterfly keyboard, touchbar. Even in the software. Look at the buggy iOS15 due to Apple being distracted on CSAM.

These are some very silly (being diplomatic here) connections being made here. Also the people that work on TV+ aren't the ones working on their chips.

 

[leman]

Yeah you can make fun of others, but at least they try and promote Apple, yes it is also for making money but still better then be negative on internet and not achieving anything. They have good videos and bad videos, no need to point out as nobody is perfect.

Wait, what? Spreading unhealthy sensationalism and misinformation and generally going a half-assed job is ok as long as one promotes something that you like? What is this, ”make Apple great again?” Apple doesn’t need that kind of disservice. They do a fairly good job, creating of irrational myths around their products does not help anyone.

 

[altaic]

Max Tech is clickbait nonsense. I commend their passion, but their content is, for instance: M1 Ultra Studio tear down— wave the thing around while yammering about irrelevant things. I had to step through frames to actually see the hardware.

For them, it’s an exposé, riding on the shoulders of iFixit (who are flawed in same same but different ways).

 

[Gerdi]

The thing is, Microsoft do not have an Arm product capable of competing with x86 product lines.

This statement is fundamentally wrong. The almost 4 years old Qualcomm 8cx compares favorably in performance against any x86 SoC in the same 7W TDP power range developed as far by Intel. This includes Amberlake-Y (2019), Lakefield (2021) and Jasperlake (2021).
So unless you compare a 7W SoC against much higher power SoCs from Intel, the 8cx Gen 1 is competing very well with at least 25% performance advantage against the closest competitor from Intel - which is much younger AND consumes almost 50% more power (N6005@10W).

 

[altaic]

This statement is fundamentally wrong. The almost 4 years old Qualcomm 8cx compares favorably in performance against any x86 SoC in the same 7W TDP power range developed as far by Intel. This includes Amberlake-Y (2019), Lakefield (2021) and Jasperlake (2021).
So unless you compare a 7W SoC against much higher power SoCs from Intel, the 8cx Gen 1 is competing very well with at least 25% performance advantage against the closest competitor from Intel - which is much younger.

How does this thread relate to your cherry-picked quote and subsequent tirade (that I happen to generally agree with). Perhaps people should try to stay on topic?

 

[Bodhitree]

This statement is fundamentally wrong. The almost 4 years old Qualcomm 8cx compares favorably in performance against any x86 SoC in the same 7W TDP power range developed as far by Intel. This includes Amberlake-Y (2019), Lakefield (2021) and Jasperlake (2021).
So unless you compare a 7W SoC against much higher power SoCs from Intel, the 8cx Gen 1 is competing very well with at least 25% performance advantage against the closest competitor from Intel - which is much younger AND consumes almost 50% more power (N6005@10W).

The power bracket doesn’t matter so much. You can find Intel laptops which offer a 15 hour battery life which stack up very well compared to Qualcomm-powered ARM devices, while offering substantially more CPU power. That is the bottom line — total system performance, not processor TDP (which is notoriously ill-defined and loosely specified anyway).

 

[Mr47]

Wait, what? Spreading unhealthy sensationalism and misinformation and generally going a half-assed job is ok as long as one promotes something that you like? What is this, ”make Apple great again?” Apple doesn’t need that kind of disservice. They do a fairly good job, creating of irrational myths around their products does not help anyone.

Well every review has it ups and downs, there is no such thing as a perfect YouTube video around. Like I said they have good point and bad points like most of them and they sometimes review stuff others completely ignore.

 

[BigPotatoLobbyist]

First of all, that Volterra PC is likely using an 8CX Gen 3 - which has four X1 cores @ 3GHz, and four A78 cores at 2.4GHz, and is fabricated on Samsung 5NM LPE, which offers dynamic power consumption higher than TSMC N7 - we can compare A78 and A710 cores on TSMC N6/N7, or Mali GPU's to Samsung 4/5NM for example, and the power suffers with Samsung's processes, even within the same manufacturer. That said, it will still offer decent or comparable performance per watt relative to what Intel and AMD have at least per core. Those X1's can realize Geekbench 1000-1100ST per leaks, and in Spec tests in phones, X1's clocked at 2.84GHz consume about 3.2-4.2W of power. The A78's would be about 65-75% of that, albeit at 2.2W in this case from the 888 where they're clocked similarly and on Samsung 5NM.

So you may be looking at a 1100/5500-6000 GB5 ST/MT device in a 15-25W profile. Not M1-tier, but it's old, reference Arm IP with just okay cache and a process that's worse than TSMC N7. I don't think it's especially good of course, but it's at least fundamentally *usable* unlike their past stuff using four A76's as their performance cores. This is probably why Microsoft are about to sell them for developers, because in 2023, Qualcomm will have TSMC-fabricated SoC's for laptops with a custom CPU architecture, but that's another story, and should prove to be Windows' "M1 moment" so to speak. And no, Apple will not be on the M3 by then, for the most part anyways (they may release the M3 on an iMac in 2023, but the small upgrade that will be the M2, M2 Pro/Max will be in 2022, 2023, and they could even skip the latter - point is - Qualcomm will be much, much closer than anyone here realizes, if not outright ahead on some metrics).

I would also note that often Intel and AMD laptops can showcase great "battery life" as defined by merely having the screen on while most of the CPU idles/gates off/low clocks and a video decode block functions, or just web browsing at 150nits with a DVFS/Windows Power setting that effectively shunts performance to dumpster time. AMD with Zen 3 actually automatically shunt performance the moment a laptop is unplugged (Changed with Rembrandt but still).


What matters isn't so much "battery life and performance" under completely distinct contexts deliberately designed to paint Intel/AMD's indices in a great light, but also battery life whilst sustaining that performance, or put another way - performance while pegged to voltages & frequencies that are favorable to power draw. I am not saying idling isn't a big deal - it absolutely is and so is acceleration for e.g. common video formats as AMD show with Rembrandt, but still. I thought people learned this with the M1, it's not just "low power" but performance at those low wattages.


Second of all: Qualcomm accelerator blocks via Hexagon are no joke. AMD and Intel may be behind on such things but Qualcomm's Hexagon actually sees use by third-party applications on Android, and they have first-party use as well as of 8Cx Gen 3. See the links below about extracting speech from audio captured - and it only requires one microphone - and wiping out background noise. Similar to RTX Voice, albeit at low power. AMD and Intel don't do much in this vein, but the idea that Qualcomm and MS are ********ting about the utility of AI acceleration with future Qualcomm chips is just lame. Don't dismiss something because you aren't familiar with it.

Voice Isolation Technology: A Feature That Speaks for Itself

developer.qualcomm.com

 

[Bodhitree]

So you may be looking at a 1100/5500-6000 GB5 ST/MT device in a 15-25W profile. Not M1-tier, but it's old, reference Arm IP with just okay cache and a process that's worse than TSMC N7. I don't think it's especially good of course, but it's at least fundamentally *usable* unlike their past stuff using four A76's as their performance cores. This is probably why Microsoft are about to sell them for developers, because in 2023, Qualcomm will have TSMC-fabricated SoC's for laptops with a custom CPU architecture, but that's another story, and should prove to be Windows' "M1 moment" so to speak.

In fact, Microsoft’s Volterra looks more capable than Apple’s Developer Preview kit. But yeah, that’s pretty much my thinking too.

And no, Apple will not be on the M3 by then, for the most part anyways (they may release the M3 on an iMac in 2023, but the small upgrade that will be the M2, M2 Pro/Max will be in 2022, 2023, and they could even skip the latter - point is - Qualcomm will be much, much closer than anyone here realizes, if not outright ahead on some metrics).

It all stands and falls by how well the Nuvia team execute their task. You could say, some of the team members have done it before, so their second iteration should be that much better. On the other hand, Apple haven’t stood still since these guys left, and they don’t have the scale of the engineering teams at Apple. Small and nimble is good for a startup, not so good for a division of a large company producing products.

I saw the marketing info they presented on AnandTech, but that was a load of braggadoccio. I think if they manage to equal the performance of the M1 they will have done well. On the other hand if Apple don’t launch a substantially improved M2, that will be them dropping the ball.

I would also note that often Intel and AMD laptops can showcase great "battery life" as defined by merely having the screen on while most of the CPU idles/gates off/low clocks and a video decode block functions, or just web browsing at 150nits with a DVFS/Windows Power setting that effectively shunts performance to dumpster time.

The problem is, people are not well educated about these things. They’re not going to stand in a store comparing performance plugged and unplugged while making up their minds to buy. Even less are they going to be aware of “snappiness” and full-stack performance.

I think in this case marketing slides are going to count for more than real performance.

 

[spiderman0616]

Apple TV shows are good. Have you seen Ted lasso and the latest dinosaur doc. I didn't find virtue signalling there.

These days, anything positive that lifts up populations of people that have been treated unequally in the past is labelled as "virtue signaling", which really just means the bullies now feel threatened.

 

[leman]

These days, anything positive that lifts up populations of people that have been treated unequally in the past is labelled as "virtue signaling", which really just means the bullies now feel threatened.

Then again, the way how Apple specifically does it can be rather awkward and artificial at times. It's not enough to have a female lead or a diverse cast, the execution matters too. For example, watching the first season of For All Mankind I had a strong impression that there was a directive to tackle as many societal issues as possible, with writers and directors totally overwhelmed and the treatment of these topics very rushed — and that's one of the better executions for an Apple TV show. Invasion takes the cake for me, with all characters being poorly written bland vanilla morons, but hey, at least there was diversity!

 

[ADGrant]

Max Tech is clickbait nonsense. I commend their passion, but their content is, for instance: M1 Ultra Studio tear down— wave the thing around while yammering about irrelevant things. I had to step through frames to actually see the hardware.

For them, it’s an exposé, riding on the shoulders of iFixit (who are flawed in same same but different ways).

The M1 Ultra Studio tear down was actually one of their better videos. But yes, most of their stuff is click bait. I would certainly not make any purchase decisions based on their recommendatons.

 

[jdb8167]

Currently, even the M1 is already 2 years old. Apple still has a lead, but it's not like everybody is resting on their laurels.

So if the M1 was say, only 17 months old, does that negate your argument? Because the M1 is not 2 years old.

 

[Freeangel1]

I enjoy Vadim's enthusiasm, but I've never seen anyone turn a termite hill into a volcano like a Max Tech video. They manage to find an obscure patent, or in this case a niche product, and make it out to be a technological revolution.

Microsoft has tried to diversify away from x86 before. Windows NT supported not just IA-32, but also PowerPC, Alpha, and MIPS. That effort went nowhere and NT 4.0 was the last to support those RISC architectures on desktop. It appears that Microsoft is hedging again with ARM, and thus far, they have achieved little with it. Many Windows programs are far too entrenched in x86, from scientific applications that require Intel-specific instructions like AVX2, to games which perform poorly, if at all, under translation. Many businesses won't even consider alternatives, simply out of habit and dependancy. Unless Intel and AMD completely collapse, I don't foresee CISC losing any appreciable Windows PC marketshare to what appears to continue to be an exclusivity deal with Qualcomm.

That may change, at some point, but Microsoft doesn't have the same structural benefits that Apple's vertical integration strategy has. As pointed out above, Microsoft doesn't control the whole stack, Apple produces the entire widget. Hence, Apple will always have an inherent advantage in that regard, including the ability to remove old technological barnacles and routinely eliminate cruft.

As long as that is the case, Microsoft will remain a reactive, over-glorified utility company, maintaining archaic compatibility until end times. It's like being impressed that my power company managed to lower my electricity bill by 2%.

over-glorified utility company


That owns 90 percent of the OS market.

which means 90 percent on ARM in the future vs APPLE's 10 percent or less

 

[Cognizant.]

over-glorified utility company


That owns 90 percent of the OS market.

which means 90 percent on ARM in the future vs APPLE's 10 percent or less

Good. Maybe Boot Camp will make a comeback.

 

[BigPotatoLobbyist]

In fact, Microsoft’s Volterra looks more capable than Apple’s Developer Preview kit. But yeah, that’s pretty much my thinking too.



It all stands and falls by how well the Nuvia team execute their task. You could say, some of the team members have done it before, so their second iteration should be that much better. On the other hand, Apple haven’t stood still since these guys left, and they don’t have the scale of the engineering teams at Apple. Small and nimble is good for a startup, not so good for a division of a large company producing products.

I saw the marketing info they presented on AnandTech, but that was a load of braggadoccio. I think if they manage to equal the performance of the M1 they will have done well. On the other hand if Apple don’t launch a substantially improved M2, that will be them dropping the ball.



The problem is, people are not well educated about these things. They’re not going to stand in a store comparing performance plugged and unplugged while making up their minds to buy. Even less are they going to be aware of “snappiness” and full-stack performance.

I think in this case marketing slides are going to count for more than real performance.

Indeed it is more powerful than the A12Z on MT in particular.


As far as Nuvia goes: I will tell you this: I’ve heard the cores are more efficient than the X3 and A720, and more efficient than Apple’s big core across the curve. People here hate hearing it but Nuvia is no joke. Andrei F works for them under Qualcomm now, too.
They also have the engineering prowess of Qualcomm behind them now. Laugh as you may, but it’s not really a startup anymore, and this greatly aids in filling in the gaps on things like SoC fabrics, modems, GPU’s… all things Qualcomm excel with.

As far as Apple goes: this forum is absolutely deluded about Apple’s near-term roadmap. We’ve already trashed the strict annual upgrade thesis on the baseline M-series chips — M2 is a 2022 product and there’s no M2 Pro/Max coming in 2022 either. M2 will almost certainly be A15 P & E cores with a clock boost from N5P. Apple already use 32MB SLC cache in the M1 and I doubt they’ll boost that to 48 or 64MB unlike the 16 -> 32MB SLC jump on the A14 to A15, which is a massive contributor to the efficiency boost. It will be LPDDR5x & some GPU boosts. Not bad but not what people think.

Qualcomm will be competing with the M2 & M3 plausibly later in 2024 for the laptops. Qualcomm will have a TSMC leading process of some kind for these chips, anything from N3 to N5, there is not going to be an easy process win for Apple here.

People keep laughing about Qc competing "with the M1". They've said *M-series*, not M1. and even then, the M2 will be a modest upgrade at best, and the M3 is a 2024 product for most Macs.

In fact, Microsoft’s Volterra looks more capable than Apple’s Developer Preview kit. But yeah, that’s pretty much my thinking too.



It all stands and falls by how well the Nuvia team execute their task. You could say, some of the team members have done it before, so their second iteration should be that much better. On the other hand, Apple haven’t stood still since these guys left, and they don’t have the scale of the engineering teams at Apple. Small and nimble is good for a startup, not so good for a division of a large company producing products.

I saw the marketing info they presented on AnandTech, but that was a load of braggadoccio. I think if they manage to equal the performance of the M1 they will have done well. On the other hand if Apple don’t launch a substantially improved M2, that will be them dropping the ball.



The problem is, people are not well educated about these things. They’re not going to stand in a store comparing performance plugged and unplugged while making up their minds to buy. Even less are they going to be aware of “snappiness” and full-stack performance.

I think in this case marketing slides are going to count for more than real performance.

RE: "People are not well-educated". This is laughably stupid and an argument against the M1's success if anything. They don't have to be educated about DVFS and power profiles of cores to intuit the performance of an Intel/AMD laptop suffers - or that performance excels while battery life suffers, for the other end of the tradeoff. They don't need to stand in a store idly and compare. Instead they'll buy and you'll read reviews about the battery life "just being ok" or performance suffering, fans blazing, and word spreads, maybe they return it, maybe consumer stereotypes develop. Weak argument.

 

[StudioMacs]

over-glorified utility company


That owns 90 percent of the OS market.

which means 90 percent on ARM in the future vs APPLE's 10 percent or less

Apple sells something north of 250 million ARM devices a year.

 

[Random_Matt]

Max is one of the most computer illiterate guys out there. And secondly, the last ARM surface device was terrible and unlikely MS will do anything good with ARM.

 

[Bodhitree]

@BigPotatoLobbyist anything that’s known about M2 let alone M3 is speculation and hearsay, just as what we know about Nuvia’s cores is pre-production rumors. It doesn’t make the slightest difference to my buying behaviour and I am certainly not going to worry about it, I’m not that invested in tech. Let’s just wait and see how it turns out.

The thing about M1’s success is, it’s been a great product for the Apple faithful, but in terms of real growth against the PC market has it been a great success? We have seen the Mac business grow by a few percent compared to the PC market’s general decline, but has M1 really moved the needle that much?