Qualcomm's M1 Rival to Be Available in PCs by Late 2023

[deconstruct60]

...
So this will be Snapdragon 8CX Generation 3
They already failed with 1 and 2 while Apple didnt with M1 because its far easier when you control the whole stack
For this to work..
...

That was a chip designed for mobile... grossly underpowered and poorly optimized for a PC replacement.

The 8CX Gen 3 is not design for mobile smartphones. There is a sizable shift here with Gen 3. Excerpt from a comparison table:

" ....

AnandTech	SD 835	SD 850	8cx Gen 1	8cx Gen 2	8cx Gen 3
Node	10LPE	10LPP	N7	N7	5nm
Prime Cores	4 x A73 2.60 GHz	4 x A75 2.95 GHz	4 x A76 2.84 GHz	4 x A76 3.15 GHz	4 x X1 3.00 GHz
Efficiency Cores	4 x A53 1.80 GHz	4 x A55 1.80 GHz	4 x A55 1.80 GHz	4 x A55 1.80 GHz	4 x A78 2.40 GHz
GPU	Adreno 540 710 MHz	Adreno 630 710 MHz	Adreno 680 585 MHz	Adreno 690 660 MHz	Adreno 8cx Gen 3

..."

Qualcomm’s 8cx Gen 3 for Notebooks, Nuvia Core in 2022/2023

www.anandtech.com

Gen 3 using mobile "big" (or at best medium size) cores as the "little cores"/E-cores. And it uses the larger for mobile X1 cores in a quad cluster. Qualcomm's and Samsung's mobile headset offerings have. one X1 core and A55 based "little cores"

https://www.anandtech.com/show/16463/snapdragon-888-vs-exynos-2100-galaxy-s21-ultra

Google's Tensor uses two X1's but still is using A55's for "little cores"

Google's Tensor inside of Pixel 6, Pixel 6 Pro: A Look into Performance & Efficiency

www.anandtech.com

The 8cx gen 3 is NOT going into any phone. There might be a larger Android Tablet with it. But there may be Windows slate Tablets too.

This will be an entirely new chip from the ground up.

Not really.

"...
IC: If you can perhaps clear something up for me: is the Nuvia team making a single core, or both a big core or a little core? Or is it that they’re dealing fully with the SoC structure into which you add in the connectivity and the graphics?

AK: It’s both, all the above. By that I mean that it's impossible for us to put out a chipset solution as sophisticated as this without having the entire system being taken into consideration. Think of it this way: the CPU by itself is part of the ‘one technology roadmap’, but so is graphics, and other things. Then we're really thinking about bringing a complete system solution to the PC and changing it in such a way that you don't go after the traditional designs. ..."
https://www.anandtech.com/show/1723...pdragon-microsoft-nuvia-and-discrete-graphics

The Nuvia cores are just one subcomponent of the die. The graphics , DSP/AI/ML cores , modem , Pluton , etc. are also components to the complete system die and will be evolutionary upgrades. More GPU cores but not necessarily a " blank, green field from ground up" GPU. Maybe some DirectX Ultimate additions but where useful on Vulkan and not a die space boat probably will see those upgrades flow though the other GPU implementations also. Same wtih DSP/AI/ML. Absolutely certainly with the modem part of the die (Windows isn't going to do much different there).

The Nuvia folks are not designing the whole chip. Qualcomm did not have to start from an empty slate to get this chip out in 2022-23. There is a more decent chance that due to the couplings between the GPU and DSP cores pefoirmance and the memory system that Qualcomm's folks are doing the memory subsystem also. All Nuvia's folks would be doing is plugging into the internal data/control bus(es) of the die. There is a very good chance Qualcomm already had (or even still has ) a team with more standard Arm X2 (or X3 depending upon timeline) designs that plug into the same internal bus .

Nuvia's cores are likely bigger so the overall die size budget probably needed to stretch but if the Nuvia acquisition didn't go through, it is extremely doubtful that Qualcomm had nothing in the hopper being developed to come out in 2023. The "tape in" of the different subsystems here is being juggled and adapted , but there likely was a foundation to build off of.



P.S. there is a pretty high likelihood that back pre-acquisition the Nuvia folks where not doing the whole server chip design either. Taking the baseline Neoverse design and just substituting in their own CPU cores would have made tons more sense with the limited time , budget, and manpower their had. Also makes them a more easier target to go an buyout of if their cores can plug into a more conventional Arm internal bus .

A system and power management , memory system , etc that already works with Linux would be a significant accelerator for adoption if most of the system die is the same.

Once they were started and had some revenue flow then taken on more internals. But if part of the objective was to get acquired ... tons of effort reinventing the wheel .. that is just a waste with little to no added value long term.

 

[Realityck]

We'll trash the competition. We are Apple fans. We won't trash the fact there is competition. Competition is important to keep everyone honest.

Funny how political types trying to make points seem to trash Apple but ignore the fact that they are offering alternatives to the competition that makes up a much larger part of the marketplace. Did the M1 just awaken all the jealousy out there when Apple leaps beyond the competition in some ways?

 

[williedigital]

Apple needs to value compatibility more IMO. Yeah it's not great to keep all those 32-bit libs around and bloat up the OS, but there should still be some way (even if much slower) to run old stuff when needed. Newer Macs can't run 32-bit Mac programs, but ironically my M1 Mac mini can run 32-bit Intel Windows programs thanks to PlayOnMac with Wine 32on64!

At big companies, all apps are going cloud. In the last 5 years I've seen opening excel become a mark of older staff not in tune with tech. Any legacy 32-bit apps can be ported to python with a skin pretty easily if they are truly that critical to the business. In general, I haven't seen any company developing new proprietary software apps--scripts and whatnot yes--but not an actual company-specific windows app.

 

[sudo-sandwich]

At big companies, all apps are going cloud. In the last 5 years I've seen opening excel become a mark of older staff not in tune with tech. Any legacy 32-bit apps can be ported to python with a skin pretty easily if they are truly that critical to the business. In general, I haven't seen any company developing new proprietary software apps--scripts and whatnot yes--but not an actual company-specific windows app.

Yeah, but at small companies (and probably some big ones), there's always those few third-party things you need to use that are old. I don't just mean 32-bit but also ones built for one macOS version ago and already broken. Creative pros tend to get hit by this the most. Also individuals, maybe even playing games. Some may say forget it, I'll use Windows and not deal with this, since the main purpose of my computer is to run the stuff I need.

Btw, if you're two OS versions behind on Mac and using Safari, even the web apps start breaking. But there are easy solutions to that.

 

[sudo-sandwich]

There are a some very important differences between then and now.

First, ARM performance is real. It's demonstrated its as faster or faster than the best Intel and AMD can produce.

Second, it's great for laptops and that it runs with low power and run cool even at high performance.

Third, the M1 Max and Ultra have shown it can scale to the highest end desktops.

Fourth, it has a high production volume that keeps it's unit costs low. In combination with iPads and iPhones, Apple is producing over 200M ARM chips a year.

The first three were all true of PPC at first. Probably the issue was the fourth one.

 

[chucker23n1]

The first three were all true of PPC at first. Probably the issue was the fourth one.

Yup. 3 million Mac sales total in 2002. Why would IBM or Motorola invest in that?

Contrast ~20 million each year today, plus another ~30 million iPads, plus another ~200 million iPhones, and all other devices using the same CPU cores. Now it suddenly makes sense.

 

[deconstruct60]

Yeah, but at small companies (and probably some big ones), there's always those few third-party things you need to use that are old. I don't just mean 32-bit but also ones built for one macOS version ago and already broken. Also individuals, maybe even playing games. Some may say forget it, let's use Windows and not deal with this.

Have to deal with it in Windows also. Windows 10 de-support is coming in 2025

Ending Support in 2025 - Microsoft Lifecycle

Find out which products will retire, reach end of support or move from mainstream support to extended support in 2025.

docs.microsoft.com

Windows 11 nukes a number of legacy 32-bit hiding placing that 'old as dirt' apps tended to live in. A 64 only kernel so 32-bit drivers are in "bad shape". DOS/16-bit legacy mode apps are toast . mainstream 32-bit apps work at the "user land" level but stuff from the previous century is likely at risk.

old school BIOS is on its way out also with Windows 11. Real UEFI without a stack of BIOS "get out jail free" cards stacked on top.

Windows 11 drew a line in the sand for x86 CPU package support also that is well past the start of the x86-64 era starting.

Windows is moving slower to dump 32-bit than Apple has , but it is moving. It is just a longer runway. Three more years and folks hiding their heads in the sand on this 32-bit issue are going to start running in walls.

Small businesses sitting on top of "dead" or "zombie" apps that have pragmatically zero real developer active support on them that date from the heydays of 32-it only wonder .... need to start to plan on when/how going to get off that train track. Deprecation Freight train is coming.

 

[sudo-sandwich]

Have to deal with it in Windows also. Windows 10 is de-support is coming in 2025

Ending Support in 2025 - Microsoft Lifecycle

Find out which products will retire, reach end of support or move from mainstream support to extended support in 2025.

docs.microsoft.com

Windows 11 nukes a number of legacy 32-bit hiding placing that 'old as dirt' apps tended to live in. A 64 only kernel so 32-bit drivers are in "bad shape". DOS/16-bit legacy mode apps are toast . mainstream 32-bit apps work at the "user land" level but stuff from the previous century is likely at risk.

old school BIOS is on its way out also with Windows 11. Real UEFI without a stack of BIOS "get out jail free" cards stacked on top.

Windows 11 drew a line in the sand for x86 CPU package support also that is well past the start of the x86-64 era starting.

Windows is moving slower to dump 32-bit than Apple has , but it is moving. It is just a longer runway. Three more years and folks hiding their heads in the sand on this 32-bit issue are going to start running in walls.

Small businesses sitting on top of "dead" or "zombie" apps that have pragmatically zero real developer active support on them that date from the heydays of 32-it only wonder .... need to start to plan on when/how going to get off that train track. Deprecation Freight train is coming.

They get a lot more time then, a duration that rivals how long many businesses last altogether. And when they're out of time, some of them will use VMs, ridiculous as it sounds. But built-in emulation could avoid that.

With a Mac it's like, every 2 years half your native software will break unless it's being actively updated, so you can never depend on a lot of third-party software.

 

[GMShadow]

Sure, but my argument is not about near-term financial outcomes. We're still in the early days of the digital age. Ford in 1922 had a much more dominant advantage in autos than Apple has now in i-devices. Somehow the market dominance of Ford behind the model T did not persist over time and Apple's won't likely either. All good for consumers.

Cheap disposable devices made to poor standards, loaded with spyware from the factory, are not “good” for anyone.

 

[deconstruct60]

Yup. 3 million Mac sales total in 2002. Why would IBM or Motorola invest in that?

Contrast ~20 million each year today, plus another ~30 million iPads, plus another ~200 million iPhones, and all other devices using the same CPU cores. Now it suddenly makes sense.

3M in 2002 would have been OK investment if. 2005 was going to be 6 or 9M. ( three year investment and 100 to 200% growth ). But it it was closer to 70%.

The other somewhat smaller problem was Apple mainly used their own "home grown" I/O chipset somewhat as a firmware dongle to keep cloning at bay. There was a big expectation at AIM start that some shared ecosystem would grow that wasn't just solely macOS as the bulk driver. Apple did exceeding little to help the non macOS parts of PPC ecosystem grow during most of that period.

Arm has enough other high volume implementations that Apple spending tons of time "moat digging" around their platform is not an impediment to the larger ecosystem. ( wasn't on x86 either when Apple went EFI when most of x86 squatted on BIOS and glacially adopted UEFI with incrementally less and less BIOS "get out jail free" cards attached. )

iPads and iPhones have large numbers but still a relatively smaller percentage of Arm ecosystem. Apple isn't buying the bulk of Arm baseline R&D expenses. It is an ecosystem where there that cost is amortized over a larger group of companies. Effectively the same with Intel. Lots of other companies besides Apple paid Intel to do the heavy lifting on CPU R&D. Apple didn't want to pay for most of it.

Finally, both Motorola and IBM spun out their foundry business ( 2004 ; Freescale (with got subsumed by NXP ) and 2014 ; Global Foundries (actually IBM paid GF to take it from them) ). Sony got IBM to do a semi-custom PPC chip for them because they paid for it. I'm sure Apple thought they were throwing decent money on the table to get the CPU processors they wanted, but I extremely doubt IBM and Motorola thought they were. Not just the low-ish volume but how much Apple was paying.

 

[seek3r]

you are very wrong. x86 / x64 are very different in design Structure and Architecture. RISC/CISC has nothing todo with Compatibility Layers, they are complete different instruction sets. The Problem with x86/64 is not the Legacy ISA, its the whole design it self. One issue is the Power Inefficiency and the huge amount of Heat Generated.

Under the hood a lot of current x86 chips are functionally RISC with a CISC emulation layer on top, so the OP’s comment you’re responding to was rather accurate

 

[seek3r]

yep as Craig Stated apple has created the way for other systems to run on M1 chips. the ball is in Microsoft's court to make a move but there stupid OEM Licensing of Windows on ARM is the only reason why there is no Windows Bootcamp on M Series Macs

Dont need to go by what they stated, current AS machines can officially boot whatever you like.

 

[Abazigal]

We'll trash the competition. We are Apple fans. We won't trash the fact there is competition. Competition is important to keep everyone honest.

There’s competition and there’s meaningful competition.

There really needs to be more distinction between the two.

 

[brand]

Until then. Real artists ship.

I think it would be more accurate to say that "real artists art". If they are not arting then they will probably be waiting tables.

 

[williedigital]

Yeah, but at small companies (and probably some big ones), there's always those few third-party things you need to use that are old. I don't just mean 32-bit but also ones built for one macOS version ago and already broken. Creative pros tend to get hit by this the most. Also individuals, maybe even playing games. Some may say forget it, I'll use Windows and not deal with this, since the main purpose of my computer is to run the stuff I need.

Btw, if you're two OS versions behind on Mac and using Safari, even the web apps start breaking. But there are easy solutions to that.

I think apple is using basic pareto thinking here. Those designers who won't update their software or customers who won't switch don't have enough value to offset the supposed gains from simplicity.

 

[williedigital]

Cheap disposable devices made to poor standards, loaded with spyware from the factory, are not “good” for anyone.

It's interesting how little of your comment has any merit.
"Cheap" --this is a positive
"disposable"--huh? iPhones are disposed of ultimately as well
"made to poor standards"--ambiguous. I assume you mean that the phones have poor build quality, but that's been debunked many times. And, of course, they can be easily and cheaply repaired compared to Apple products.
"loaded with spyware"--old person thinking. The market has shown over and over and over that privacy is not a core concern in tech. We can look forward to a world where almost nothing about us is secret, and we will be happier for it.

 

[chucker23n1]

"loaded with spyware"--old person thinking. The market has shown over and over and over that privacy is not a core concern in tech. We can look forward to a world where almost nothing about us is secret, and we will be happier for it.

Ah yes, please do keep explaining to others that something that doesn't concern you shouldn't concern anyone else either.

 

[Abazigal]

It's interesting how little of your comment has any merit.
"Cheap" --this is a positive
"disposable"--huh? iPhones are disposed of ultimately as well
"made to poor standards"--ambiguous. I assume you mean that the phones have poor build quality, but that's been debunked many times. And, of course, they can be easily and cheaply repaired compared to Apple products.
"loaded with spyware"--old person thinking. The market has shown over and over and over that privacy is not a core concern in tech. We can look forward to a world where almost nothing about us is secret, and we will be happier for it.

I guess instead of “cheap”, we should use the term “lower value”.

An item can be cheaper, but if it lasts shorter and spoils faster and you find yourself having to replace it more often, it can end up costing more in the long run compared to if you had paid more for a higher quality alternative.

And the cheaper a phone is, the less likely people will pay to have it repaired, since they can easily get a replacement for just a little more cash (and they usually use it as an excuse to upgrade anyways, especially if your phone is no longer receiving software updates).

Ironically, the iphone market share continues to grow in part because of how long-lived iPhones are. They are typically supported for 4-5 years at least, have decent build quality, and typically command more on the 2nd hand market.

 

[OnawaAfrica]

Dont need to go by what they stated, current AS machines can officially boot whatever you like.

yes i know the Bootloader on M Series Macs is open. but still if there are no way by Microsoft to install windows on ARM on a machine then i can say as many time as u want the loader is open.

Microsoft has to provide a way to isntall it like u would on a x86 pc but they refuse and they only want to give lciense to OEM so microsoft is the issue here

 

[chucker23n1]

Is that really in Microsoft's interest to do so?

I think neither Apple nor Microsoft particularly care to invest much into having Windows run on ARM Macs, but also, it's not against either party's interest. It just isn't a particularly high-priority pursuit.

(Why would it be against Microsoft's interest to have Windows run on more machines?)

 

[macsplusmacs]

I think it would be more accurate to say that "real artists art". If they are not arting then they will probably be waiting tables.

True. But the version I posted is a famous quote from Steve Jobs.

 

[Unregistered 4U]

It really doesn't even matter if it leapfrogs Apple Silicon. We always put so much focus on the ceiling of performance that we fail to notice how much the floor has been raised. The cheapest Mac you can buy today is a beast compared to the cheapest PC you can buy. Today, in 2022, you can buy a Core i3 HP all-in-one for under $400 that barely has enough horsepower to keep up with Windows Update and was probably already out of date 5 years before it was designed. A bare bones M1 MacBook Air will probably still be running whatever you throw at in 5+ years from now.

In five years, if the next "killer app" is something that requires decent CPU/GPU power (think AR/VR or something like that) a lot of Mac users will be potential customers even on several years old off-the-shelf computers.

Here’s one of those kinds of things (a Monterey feature) that can be done all the way down on the lowest level M1, but requires 16GB of VRAM on Intel. It’s already started.
Object Capture
Turn a series of 2D images into a photo-realistic 3D object that’s optimized for AR in just minutes using the power of Mac. Object Capture makes 3D content creation easy for all developers.

 

[Unregistered 4U]

That's not strictly true. Apple doesn't compete among commodity PC makers, but there is still competition at the system level, just like there's competition among smartphone vendors.

It’s competition, but not “competition is good” competition. Qualcomm could come out with a processor 3 times as fast as Apple Silicon, but, as it doesn’t run macOS and can’t run macOS it can’t “compete” against anything for which “running macOS” is important to a consumer. There’s nothing about that “competition” that will drive Apple to do anything different from their current roadmap. Apple’s processor performance will increase on the timetable they’ve already set out, regardless of what Intel, AMD, or Qualcomm is doing.

 

[Unregistered 4U]

Apple friends-don’t sleep on what Qualcomm is doing. They are a major company that has lots of resources

..and zero capability to run macOS.

Even if Windows/Qualcomm is an amazingly wonderful product, all it’ll mean is that the number of systems every year sold that are NOT-Mac will alter their CPU pie chart a bit.

 

[Unregistered 4U]

Apple could use the competition. I was surprised to learn that the latest MSI flagship laptop destroys the MBP Max in performance. Sure, it runs hot and the battery lasts half as long, but the Max runs hot too and most people I know run their laptops plugged in when they’re doing heavy duty work.

There will always be “something” that destroys any MBP Max in performance in any given year. That’s guaranteed. And, it’s not any kind of competition that’s going to drive Apple to do anything different than they’re already doing. Because, if someone wants to run macOS and macOS applications, then no matter how seriously spanked the MBP Max got, they’re going to buy that MBP Max.