Discussion Apple Silicon SoC thread

[Eug]

M1
5 nm
Unified memory architecture - LP-DDR4
16 billion transistors

8-core CPU

4 high-performance cores
192 KB instruction cache
128 KB data cache
Shared 12 MB L2 cache

4 high-efficiency cores
128 KB instruction cache
64 KB data cache
Shared 4 MB L2 cache
(Apple claims the 4 high-effiency cores alone perform like a dual-core Intel MacBook Air)

8-core iGPU (but there is a 7-core variant, likely with one inactive core)
128 execution units
Up to 24576 concurrent threads
2.6 Teraflops
82 Gigatexels/s
41 gigapixels/s

16-core neural engine
Secure Enclave
USB 4

Products:
$999 ($899 edu) 13" MacBook Air (fanless) - 18 hour video playback battery life
$699 Mac mini (with fan)
$1299 ($1199 edu) 13" MacBook Pro (with fan) - 20 hour video playback battery life

Memory options 8 GB and 16 GB. No 32 GB option (unless you go Intel).

It should be noted that the M1 chip in these three Macs is the same (aside from GPU core number). Basically, Apple is taking the same approach which these chips as they do the iPhones and iPads. Just one SKU (excluding the X variants), which is the same across all iDevices (aside from maybe slight clock speed differences occasionally).

EDIT:



M1 Pro 8-core CPU (6+2), 14-core GPU
M1 Pro 10-core CPU (8+2), 14-core GPU
M1 Pro 10-core CPU (8+2), 16-core GPU
M1 Max 10-core CPU (8+2), 24-core GPU
M1 Max 10-core CPU (8+2), 32-core GPU

M1 Pro and M1 Max discussion here:

Page 78 - Discussion - Apple Silicon SoC thread

Page 78 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

M1 Ultra discussion here:

Page 109 - Discussion - Apple Silicon SoC thread

Page 109 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

M2 discussion here:

Page 127 - Discussion - Apple Silicon SoC thread

Page 127 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

Second Generation 5 nm
Unified memory architecture - LPDDR5, up to 24 GB and 100 GB/s
20 billion transistors

8-core CPU

4 high-performance cores
192 KB instruction cache
128 KB data cache
Shared 16 MB L2 cache

4 high-efficiency cores
128 KB instruction cache
64 KB data cache
Shared 4 MB L2 cache

10-core iGPU (but there is an 8-core variant)
3.6 Teraflops

16-core neural engine
Secure Enclave
USB 4

Hardware acceleration for 8K h.264, h.264, ProRes

M3 Family discussion here:

Page 215 - Discussion - Apple Silicon SoC thread

Page 215 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

M4 Family discussion here:

Page 263 - Discussion - Apple Silicon SoC thread

Page 263 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

 

[moinmoin]

There were still newer, slightly better chips. Apple was in no hurry.

Apple had no control over quality, quantity and timing (and Intel botched a lot there over the years). Now it does.

Just imagine Apple being in the position of Nvidia, announcing that its latest and greatest product is using SPR, only for Intel to delay it once more hours later.

 

[repoman27]

There were still newer, slightly better chips. Apple was in no hurry.

Have you ever bothered to line up the availability dates of all of the components in each Mac model with its release date to see if it actually could have shipped sooner? There were occasions when Apple had to delay the introduction of a new Mac due to issues that were not related to component supply, but they were incredibly rare. There were also several instances where Apple wasn't buying what the vendors were selling and chose to skip a generation.

Releasing new product as soon as possible in the component life cycle maximizes sales and profit. Do you really think Tim Cook's Apple is in no hurry to maximize profit when they otherwise could? It's not like they're hanging out for a quarter while they draw down channel inventory.

 

[poke01]

Dell's 2022 XPS is bad compared to the M2 MacBook Air.

Only saving grace is it's $999 vs $1199.

The new XPS has no 3.5mm jack and has worse screen(lower brightness, lower screen res) and 720p. Less ports than M2 Air and likely get loud.

The M2 MacBook Air has it beat in every hardware spec but external monitor support.

I would go for the M2 Air.

Dell XPS 13 (2022): Release date, price, and everything you need to know

Dell has introduced the 2022 model of its iconic XPS 13, and there's a lot that's changed. Here's everything you need to know.

www.xda-developers.com

MacBook Air 13- and 15-inch with M3 or M2 - Tech Specs

See all technical specifications for the 13- and 15-inch MacBook Air with M3 chip and the 13-inch MacBook Air with M2 chip.

www.apple.com

 

[gdansk]

Dell's 2022 XPS is bad compared to the M2 MacBook Air.

Only saving grace is it's $999 vs $1199.

The new XPS has no 3.5mm jack and has worse screen(lower brightness, lower screen res) and 720p. Less ports than M2 Air and likely get loud.

The M2 MacBook Air has it beat in every hardware spec but external monitor support.

I would go for the M2 Air.

Dell XPS 13 (2022): Release date, price, and everything you need to know

Dell has introduced the 2022 model of its iconic XPS 13, and there's a lot that's changed. Here's everything you need to know.

www.xda-developers.com

MacBook Air 13- and 15-inch with M3 or M2 - Tech Specs

See all technical specifications for the 13- and 15-inch MacBook Air with M3 chip and the 13-inch MacBook Air with M2 chip.

www.apple.com

Dell has two XPS 13 models this year. The Plus has a great OLED 3456x2160 screen.
The price, processor and lack of headphones port is a downgrade, however.

 

[Doug S]

A = 12 months

M = 18 months​

M Pro/Max/Ultra = 24 months​

M Mega = 30 months​

Are you saying you think the Mac Pro is going to have its own die, and not be built as 4x Max SoCs like the Ultra is built with 2x Max SoCs? Even though Apple patents showing I/O pads for three off chip connections have been posted?

If there's a different cadence for the Pro/Max die, there will be the exact same cadence for Ultra and Extreme or Mega or whatever you want to call what goes in the Mac Pro. And if there's a different cadence for the basic Mx and Pro/Max, it will be because they update Mx every year just like Ax.

 

[JasonLD]

Just imagine Apple being in the position of Nvidia, announcing that its latest and greatest product is using SPR, only for Intel to delay it once more hours later.

Well, I am sure Nvidia knew about it before Intel made SPR's delay known to the public. Most likely SPR delay doesn't impact them.

 

[repoman27]

Are you saying you think the Mac Pro is going to have its own die, and not be built as 4x Max SoCs like the Ultra is built with 2x Max SoCs? Even though Apple patents showing I/O pads for three off chip connections have been posted?

If there's a different cadence for the Pro/Max die, there will be the exact same cadence for Ultra and Extreme or Mega or whatever you want to call what goes in the Mac Pro. And if there's a different cadence for the basic Mx and Pro/Max, it will be because they update Mx every year just like Ax.

The existing M1 Max die cannot be used to make a 4-die SoC. A 4-up design like SPR will require two additional layouts / dies (right and left mirrored) with additional D2D interfaces. So yes, I'm saying that the Mac Pro chip will have a different cadence, as is already clearly the case. Apple has publicly stated that a Mac Pro is forthcoming, but the SoC will not be part of the M1 family. They did two big M1 layouts (one full layout, plus a chop), requiring two tape-outs and mask sets, which were packaged three different ways to make three very different SoCs (M1 Pro, M1 Max, and the multi-die M1 Ultra). There is no guarantee that they will continue to pursue that same strategy going forward, let alone with each generation of cores. I'm saying that I think they'll do the 4-up "Mega" dies for the M2 generation.

FWIW, Gurman reported this for Bloomberg back on April 14 of this year:

The new machines being tested include:

• A MacBook Air with an M2 chip, codenamed J413. This Mac will have eight CPU cores, the components that handle the main processing, and 10 cores for graphics. That’s up from eight graphics cores in the current MacBook Air.
• A Mac mini with an M2 chip, codenamed J473. This machine will have the same specifications as the MacBook Air. There’s also an “M2 Pro” variation, codenamed J474, in testing.
• An entry-level MacBook Pro with an M2 chip, codenamed J493. This too will have the same specifications as the MacBook Air.
• A 14-inch MacBook Pro with M2 Pro and “M2 Max” chips, codenamed J414. The M2 Max chip has 12 CPU cores and 38 graphics cores, up from 10 CPU cores and 32 graphics cores in the current model, according to the logs. It will also have 64 gigabytes of memory.
• A 16-inch MacBook Pro with M2 Pro and M2 Max chips, codenamed J416. The 16-inch MacBook Pro’s M2 Max will have the same specifications as the 14-inch MacBook Pro version.
• A Mac Pro, codenamed J180. This machine will include a successor to the M1 Ultra chip used in the Mac Studio computer.
• Apple is also testing a Mac mini with an M1 Pro chip, the same processor used in the entry-level 14-inch and 16-inch MacBook Pros today. That machine is codenamed J374. The company has tested an M1 Max version of the Mac mini as well, but the new Mac Studio may make these machines redundant.

The new MacBook Air, low-end MacBook Pro and new Mac mini are scheduled to debut as early as this year, with at least two Macs planned for launch around the middle of the year, Bloomberg has previously reported. The new MacBook Air is destined to be the product’s biggest redesign in its history, adding a thinner frame and MagSafe charging.

And this yesterday:

Apple is also planning new high-end MacBook Pros with M2 Pro and M2 Max chips for as early as the end of 2022, though the release date may slip into early 2023. These new 14-inch and 16-inch models, codenamed J414 and J416, won’t be radically new products beyond offering the speedier chips. Apple is also releasing a new 13-inch MacBook Pro next month with the same M2 chip as the MacBook Air.

The M2 Max chip in the next high-end MacBook Pros will include 12 main processing cores and up to 38 graphics cores, up from a maximum of 10 processing cores and 32 graphics cores in the current models. New versions of the Mac mini and a revamped Mac Pro are also in testing within Apple, Bloomberg has reported. And the company is already working on an M3 chip destined for a future iMac and other products.

 

[moinmoin]

Well, I am sure Nvidia knew about it before Intel made SPR's delay known to the public. Most likely SPR delay doesn't impact them.

It's a bad look either way, and it's a public example. We can expect most such examples never making it to the public for obvious reasons. Point is Apple is better off alone and Macs might well see more regular refreshes now without Intel's node shackles.

 

[Heartbreaker]

At right around 40 million units for basic M series devices (M1, M2, etc.), annual cadence isn't crazy at all. It actually seems somewhat likely.

At that scale it really doesn't make sense to go annually. All the trends going forward signal longer intervals rather than shorter ones.

Examine Ax performance delta generation over generation. It's trending strongly down. Which means with smaller gains/generation, it makes more sense to skip a generation going forward to have a more tangible reason to upgrade. Rather than just do a annual upgrades for the sake of novelty.

Especially when combined with ever rising tape-out costs. You really don't wan't to incur a new tapeout unless you have to.

They can start to align new chips with the occasional big improvements in underlying cores, which won't be annually. Or if there are no big improvements in cores, until they accumulate enough smaller gains.

As I said in a previous post. Once the transition is complete, I expect the pace will relax further, and there will be new chips at intervals convenient to Apple, not slavishly following the calendar, nor pleasing forum posters who want something new every week to talk about.

I expect a completely irregular M series upgrade schedule, but most definitely not an annual one.

 

[Eug]

Apple M2 Die Shot and Architecture Analysis – Big Cost Increase And A15 Based IP

Apple announced their new 20 billion transistor M2 SoC at WWDC. Unfortunately, it’s quite a minor uplift in performance in some areas such as CPU. Apple’s gains mostly came from the GPU and video editing side of things.

semianalysis.substack.com

 

[Roland00Address]

*Grumbles* about the lack of a 2nd display controller. The die size is not even 150mm2, one can spend a few extra mm2 on the die 😛

The lack of reviews for a month is going to make this space more negative than positive till this is rectified.

 

[Doug S]

One thing I've been interested to see was whether Apple rectified the bandwidth mismatch between L2/SLC and LPDDR5 to improve scaling. Semianalysis' article seems to indicate they have addressed that with M2, so it will be interesting to see how the usable bandwidth per core and for all cores compares between M1 and M2. That won't affect performance much for a lot of stuff, but where it matters it can really make a difference.

It should also improve scaling for the GPU - at the very least to match the improvement in memory bandwidth if they move to LPDDR5X in the Pro/Max (like how they got faster RAM than M1) and 25% more GPU cores, though ideally would go beyond that so closer to the full bandwidth can be exploited by the GPU.

 

[repoman27]

At that scale it really doesn't make sense to go annually. All the trends going forward signal longer intervals rather than shorter ones.

Apple already updates their microarchitectures on an annual cadence and is driving TSMC to deliver incremental process updates each year as well. The basic M series SoC is just a slightly larger A series with a few additional cores and I/O blocks. Most of the hard work has already been done and paid for, so there is very little reason not to update the Mx annually.

If Apple spent $550 million designing the Mx (which is unlikely because it isn't even close to a clean sheet design) and can sell 40 million of them a year, that works out to a design cost of $13.75 per unit. If Apple could match that sales pace over a two year cycle (which is also unlikely) the design cost per unit would go down by less than $6.88. Apple is far more interested in growing their installed base and adding new users, which is is a lever for recurring services revenue, then they are in recovering every last penny of gross margin on hardware.

The annual sales of M series chips is significantly higher than any recent AX series, and higher than many client oriented dies produced by Intel, AMD, and NVIDIA.

The average refresh interval for AX / M series is just under 18 months, with a standard deviation of right around 6 months. In other words, the lower bound we should expect for M series releases is 12 months and the longest we would anticipate is 24 months. An irregular release cadence is not beneficial to Apple or their customers, and not all quarters are created equal. If Apple were shooting for 18 months between releases, they would probably have to settle on a mid-November / mid-May launch cycle. The first M1 Macs went on sale Nov 17, 2020. The first M2 Macs will be available in July 2022 but were announced in the first week of June. Given extenuating circumstances such as a global pandemic, supply chain disruption, war in Europe, inflation, etc., Apple was still less than two months off of delivering the first refresh on an 18-month schedule.

Examine Ax performance delta generation over generation. It's trending strongly down. Which means with smaller gains/generation, it makes more sense to skip a generation going forward to have a more tangible reason to upgrade. Rather than just do a annual upgrades for the sake of novelty.

Especially when combined with ever rising tape-out costs. You really don't wan't to incur a new tapeout unless you have to.

They can start to align new chips with the occasional big improvements in underlying cores, which won't be annually. Or if there are no big improvements in cores, until they accumulate enough smaller gains.

As I said in a previous post. Once the transition is complete, I expect the pace will relax further, and there will be new chips at intervals convenient to Apple, not slavishly following the calendar, nor pleasing forum posters who want something new every week to talk about.

I expect a completely irregular M series upgrade schedule, but most definitely not an annual one.

Apple releases devices in new colors just to kick sales in subsequent quarters once supply balance is achieved. Novelty isn't a bad thing if it works. I'm pretty sure Apple will do whatever is required to maximize long term growth. Right now, they are behind, and they need to accelerate their pace if they want to catch up to and overtake Intel / AMD / NVIDIA. I get the distinct impression that Johny Srouji would very much like Apple to have an unquestioned performance lead in the client PC space, just as they currently do for mobile SoCs.

 

[Mopetar]

I don't think the design cost for a specific chip is a good way to look at it. Apple's design team is costing them a roughly fixed amount per year regardless of what they're working on. More time on one chip is just time not working on a successor. If the financials didn't make sense for them they would have given up long ago.

Trying to look at the costs in that way is too short-sighted and is what has lead a lot of companies into making foolish decisions just to try to have one good quarter. Apple has shown that they don't think of things that way. If it's not something that they think works out, we probably don't even hear about it.

 

[Doug S]

I don't think the design cost for a specific chip is a good way to look at it. Apple's design team is costing them a roughly fixed amount per year regardless of what they're working on. More time on one chip is just time not working on a successor. If the financials didn't make sense for them they would have given up long ago.

Trying to look at the costs in that way is too short-sighted and is what has lead a lot of companies into making foolish decisions just to try to have one good quarter. Apple has shown that they don't think of things that way. If it's not something that they think works out, we probably don't even hear about it.

You say "design team" singular assuming there's just one. That's highly unlikely to be the case - and if it is they'll have multiple subteams (which may have some people on more than one) in that one all-encompassing team, which is basically the same thing.

You're right that salaries are paid either way but unless they abandon their habit of doing a new SoC for each new iPhone they will continue to have new cores available every year. Why not use them? If they reuse everything that's not in Ax (i.e. the interchip glue, TB controllers, whatever) every other year the cost is mostly the floorplan and routing, and mask sets. You saying Apple will "think different" and decide that's not an efficient use of engineering resources?

18 months seems like a terrible cadence, because it is a 1 year / 2 year gap technology wise - i.e. if M3 is 18 months after M2 it will be based on A17 cores two years more advanced than M2's A15 cores.

24 months only makes sense if Apple also goes to 24 months with iPhone, perhaps doing something akin to Intel's tick tock model where every other year is mostly just a tweak and shrink. They've extracted the "easy" IPC, maybe at some point based on diminishing returns they'll decide it just isn't worth doing new cores every year any longer, and they would rather give themselves more time to do bigger changes.

I just have a really hard time seeing how it makes sense for Apple to settle on something other than a yearly cadence for M (after the ramp up is complete with the release of the Mac Pro) so long as they are doing A yearly.

 

[igor_kavinski]

Apple M1 Affected By "PACMAN" Hardware Vulnerability In Arm Pointer Authentication - Phoronix

www.phoronix.com

Researchers from MIT found that the Arm Pointer Authentication functionality within the M1 can be defeated and without traces. The researchers allege, "PACMAN utilizes a hardware mechanism, so no software patch can ever fix it."

Any Apple developer wanna chime in on how serious this could be? It's likely not fixed in M2 since it's so new.

 

[mikegg]

Apple already updates their microarchitectures on an annual cadence and is driving TSMC to deliver incremental process updates each year as well. The basic M series SoC is just a slightly larger A series with a few additional cores and I/O blocks. Most of the hard work has already been done and paid for, so there is very little reason not to update the Mx annually.

If Apple spent $550 million designing the Mx (which is unlikely because it isn't even close to a clean sheet design) and can sell 40 million of them a year, that works out to a design cost of $13.75 per unit. If Apple could match that sales pace over a two year cycle (which is also unlikely) the design cost per unit would go down by less than $6.88. Apple is far more interested in growing their installed base and adding new users, which is is a lever for recurring services revenue, then they are in recovering every last penny of gross margin on hardware.

The annual sales of M series chips is significantly higher than any recent AX series, and higher than many client oriented dies produced by Intel, AMD, and NVIDIA.

The average refresh interval for AX / M series is just under 18 months, with a standard deviation of right around 6 months. In other words, the lower bound we should expect for M series releases is 12 months and the longest we would anticipate is 24 months. An irregular release cadence is not beneficial to Apple or their customers, and not all quarters are created equal. If Apple were shooting for 18 months between releases, they would probably have to settle on a mid-November / mid-May launch cycle. The first M1 Macs went on sale Nov 17, 2020. The first M2 Macs will be available in July 2022 but were announced in the first week of June. Given extenuating circumstances such as a global pandemic, supply chain disruption, war in Europe, inflation, etc., Apple was still less than two months off of delivering the first refresh on an 18-month schedule.


Apple releases devices in new colors just to kick sales in subsequent quarters once supply balance is achieved. Novelty isn't a bad thing if it works. I'm pretty sure Apple will do whatever is required to maximize long term growth. Right now, they are behind, and they need to accelerate their pace if they want to catch up to and overtake Intel / AMD / NVIDIA. I get the distinct impression that Johny Srouji would very much like Apple to have an unquestioned performance lead in the client PC space, just as they currently do for mobile SoCs.

Exactly.

Not sure what is so hard for people to understand here.

Most reliable leakers said the M2 was ready by the Spring Event (March 2022). In fact, reliable Mark Gurman even said he expected the M2 to launch with the 13" MBP in the Spring Event because the Air wasn't ready. But obviously Apple would rather launch the M2 with the new Air design together. It would have been strange to launch the M2 in Apple's completely outdated design.

Supply chain issues, work from home, inflation, and complexity with bundling a new Air design with a new SoC meant that we shouldn't rely on the 18-month cadence.

The mere fact that Apple decided to use the A15 cores would suggest that they want to do annual updates.

And you're right that the A Series is not too different from the base M series. It just has more CPU and GPU cores. All the blocks are already there. The bulk of the design work was already done. Same CPU cores, GPU cores, Neural Engine, encode/decode blocks.

And finally, once again, the base M goes into Macbook Air, iPad Air, iPad Pro 13, iPad Pro 11, Mac Mini, iMac, 13" MBP. Rumors are that a 15" Air is coming next year too. And a new 12" Macbook. The number of devices relying on the base M is staggering. Apple is really getting their value.

It makes no sense for Apple to only update so many devices once every 2 years. Apple didn't work so hard to transition the Mac away from Intel just so they could update the chip less often.

Pro/Max/Ultra/"Extreme", I don't expect annual updates. The base M, I absolutely do.

 

[Heartbreaker]

Exactly.

Not sure what is so hard for people to understand here.

Most reliable leakers said the M2 was ready by the Spring Event (March 2022).

Sorry, but reality is more reliable than leakers. It's absurd to talk about leaks mattering more than what actually happened.

 

[Heartbreaker]

Apple already updates their microarchitectures on an annual cadence and is driving TSMC to deliver incremental process updates each year as well. The basic M series SoC is just a slightly larger A series with a few additional cores and I/O blocks. Most of the hard work has already been done and paid for, so there is very little reason not to update the Mx annually.

They have already demonstrated they are NOT doing annual updates. They will do them when convenient. Convenience will lead to a variable schedule, just like they were for the Ax chips.

Developing a new core is hard work. But that doesn't lessen the sting of new tapeouts.

A lot of people don't realize, that this was almost the entire point of Ryzen MCM and Chiplets. To do one tapeout and use it across a range of products. It's mostly about minimizing new design tapeouts.

Every tapeout saved is potentially hundred million+ saved. No one should burn 100 million without VERY good reason.

Apple releases devices in new colors just to kick sales in subsequent quarters once supply balance is achieved. Novelty isn't a bad thing if it works.

That isn't an argument for annual SoC updates, in fact it highlights the opposite case. Normal people probably care more about colors for novelty, that about novelty in the SoC, so spending hundred million+ on a new SoC just for the sake of novelty isn't sound business, when you can just get it from pennies spent on a different color.

Ultimately, arguing for annual M chips, seems like a late and lost argument at this point. Apple has already demonstrated the cadence will be at their convenience, and that convenience will be irregular.

People on forums like neat and tidy stories, like a regular annual, 18 month, 24 month cadence, but reality is usually much messier.

 

[repoman27]

They have already demonstrated they are NOT doing annual updates. They will do them when convenient. Convenience will lead to a variable schedule, just like they were for the Ax chips.

Developing a new core is hard work. But that doesn't lessen the sting of new tapeouts.

A lot of people don't realize, that this was almost the entire point of Ryzen MCM and Chiplets. To do one tapeout and use it across a range of products. It's mostly about minimizing new design tapeouts.

Every tapeout saved is potentially hundred million+ saved. No one should burn 100 million without VERY good reason.

That isn't an argument for annual SoC updates, in fact it highlights the opposite case. Normal people probably care more about colors for novelty, that about novelty in the SoC, so spending hundred million+ on a new SoC just for the sake of novelty isn't sound business, when you can just get it from pennies spent on a different color.

Ultimately, arguing for annual M chips, seems like a late and lost argument at this point. Apple has already demonstrated the cadence will be at their convenience, and that convenience will be irregular.

People on forums like neat and tidy stories, like a regular annual, 18 month, 24 month cadence, but reality is usually much messier.

I showed you the numbers. There are very strong arguments that:

a.) That the lower bound for a regular interval between releases is 12 months.
b.) That Apple is likely targeting an 18 month refresh cycle.
c.) The cadence for the entry level M series SoCs will be less variable than the AX series was.
d.) That there are compelling reasons, both financial and strategic, to accelerate releases to 12 months in the near future.

You have presented zero evidence and are instead insisting that your gut feelings (as a poster on an Internet forum) should triumph over rational discussion by others of publicly available information. Many of your assertions are dismissals of valid data points that have been presented. I try my best not to be dogmatic, but your arguments don't align particularly well with the facts on the ground.

Being a poster on an Internet forum myself, I realize that my predictions have no bearing on Apple's future release schedule. However, I am more than a casual observer. I have been following apple hardware closely for nearly 40 years at this point, and started working on a financial analysis of Apple developing their own SoCs for the Mac over four years ago. I'm pretty sure your model isn't accounting for as many variables as mine is.

 

[Doug S]

Apple M1 Affected By "PACMAN" Hardware Vulnerability In Arm Pointer Authentication - Phoronix

www.phoronix.com

Any Apple developer wanna chime in on how serious this could be? It's likely not fixed in M2 since it's so new.

It is a flaw in a security feature, not an exploit itself. It is like if you added a fence around your house as an added defense against burglars, which was found to be useless against thieves who bring a ladder. It doesn't make the lock on your front door or motion sensors inside the house stop working.

 

[igor_kavinski]

It is a flaw in a security feature, not an exploit itself.

How severe is this flaw? Severe enough that if an exploit is developed, Apple is kaput?

 

[Mopetar]

How severe is this flaw? Severe enough that if an exploit is developed, Apple is kaput?

It's more benign than you might expect. It just means that a certain class of vulnerabilities which either do or don't exist which could be stopped by this particular security mechanism would now be able to bypass it in a way that would not be detectable.

This doesn't expose any kind of new attack vector that could open up the hardware to be taken over by some exploit. Like a previous poster said, this only nullifies a security feature.

 

[Doug S]

How severe is this flaw? Severe enough that if an exploit is developed, Apple is kaput?

What in the world are you talking about? This is ADDED protection, which most CPUs (including Intel's and AMD's) do not have and they are not "kaput". Do you think Apple abandoned all other layers of security to rely solely on PAC?

If you put a fence around your house for security, if the fence is breached you are in no worse condition than you were before the fence was added. Unless you are really dumb and leave your front wide open all the time because you think that fence will protect you. It seems that's what you are suggesting Apple did when they added PAC.

 

[Heartbreaker]

You have presented zero evidence and are instead insisting that your gut feelings (as a poster on an Internet forum) should triumph over rational discussion by others of publicly available information. Many of your assertions are dismissals of valid data points that have been presented. I try my best not to be dogmatic, but your arguments don't align particularly well with the facts on the ground.

You have also presented ZERO Evidence. Just speculation. Your speculation isn't superior just because you think it is.

The ONLY cadence data-point we actually have is a 19 month gap between M1 and M2. That isn't indicative of an annual trend. Not even close.

That is the most significant fact. There is an argument for ~18 month schedule or maybe a variable one driven by Apples convenience, but NOT an annual one.

You've just become enamored with the idea of an annual release, and are rationalizing.