Looks like Apple is on a 2-year update cycle for AS Macs. If so, I think Intel will fly past them.

[ultrakyo]

It potentially be yearly after completion of AS transition since now they control control the M series roadmap instead of relying on intel.

Apple religious event calendar should be like this:

March/April - Special Event (if any)
June - WWDC
September - A series Chip & Hardware
October - M Series Chip & Hardware

So we may see Mac Pro AS in March/April event at the earliest.

 

[SirSpuitElf]

Rumor is the M2 won’t be released until possibly late next year. With the M2 Pro/Max in 2023.

If this is the case, I absolutely see Intel flying right past Apple sooner rather than later.

Intel’s generational improvement will be less, but they will be updating chips twice as often.

I honestly think Intel will eventually blow our mind. They’re pushing the limits of older technologies and when they start using newer technologies, I have to see what they will create.

 

[Fomalhaut]

I honestly think Intel will eventually blow our mind. They’re pushing the limits of older technologies and when they start using newer technologies, I have to see what the will create.

This is good news; Having serious competitors pushing the boundaries will prevent Apple from stagnating or sand-bagging their progress. It can only benefit consumers.

 

[LIVEFRMNYC]

Apple and AMD have been showing me balance along with great performance. This is why I have a Ryzen 9 3950 that's easy to keep temps under control in my desktop, a MBA M1, and a MBP 16' M1 Max on preorder. When Intel shows the same, then I will pay attention to them again.

 

[Ethosik]

Raw CPU speeds are no longer a concern. It doesn't matter if Intel will have more cores or performance. People are still thinking these M1 and Pro/Max are simple CPUs. What they don't realize we have GPU, encode/decode codecs, separate neural cores for ML tasks, shared memory, faster memory than the exchange from CPU to GPU on a PCIe x16 lane then back to CPU and back to memory. This all adds up. It is NOT just about the CPU portion and people need to realize this.

I can get the best Intel processor, but stick with standard SATA SSDs. Therefore, I will lose out of the 7 GB/s SSD speeds that Apple's ENTIRE CHIP can do. I can get the best Intel processor but stick an old GTX 1060 in there where Apple's GPU might beat that like crazy.

 

[Ethosik]

Devils advocate. What’s the downside of intel doing well?

If they somehow manage to outperform internal silicon there isn’t much stopping apple from going back?

It’s not about intel or AS… Competition is a good thing, everyone wins.

Yep. Competition is good. Also, I still use Windows systems. As someone that has used AMD systems, even recently, they still have a long way to go with the compatibility and reliability (USB issues and things like that). So I still prefer Intel on Windows based systems.

 

[bousozoku]

Yes it does. Macs don't exist in a vacuum. Today, more Windows laptops dwarf the marketshare of Macs. Apple should be aggressive in absolutely crushing the computer market within the next 5 years.

Apple is concerned about profit, not about having the biggest market share. In the 1980s, they had multiple chances to capture most everything, but they chose to litigate, rather than innovate. In the 1990s, that previous strategy could have finished off the company. Ellen Hancock was probably the main reason Apple made it to all the good things that happened from 2000 onward.

 

[Ethosik]

Apple is concerned about profit, not about having the biggest market share. In the 1980s, they had multiple chances to capture most everything, but they chose to litigate, rather than innovate. In the 1990s, that previous strategy could have finished off the company. Ellen Hancock was probably the main reason Apple made it to all the good things that happened from 2000 onward.

Reminds me of this:

And Steve is right. While there are $300 Dell systems, I would not recommend people buy them.

 

[alien3dx]

It potentially be yearly after completion of AS transition since now they control control the M series roadmap instead of relying on intel.

Apple religious event calendar should be like this:

March/April - Special Event (if any)
June - WWDC
September - A series Chip & Hardware
October - M Series Chip & Hardware

So we may see Mac Pro AS in March/April event at the earliest.

i just want to wait till they said. 1 nanometer chip .the last my friend said in taiwan max 2 or 1 not remember. now apple 5 and intel higher.Which mean intel hang long way.But if apple stuck those event would be troublesome of they make wafer bigger and bigger ...to support those event..hard moved actually.

 

[Fomalhaut]

Raw CPU speeds are no longer a concern. It doesn't matter if Intel will have more cores or performance. People are still thinking these M1 and Pro/Max are simple CPUs. What they don't realize we have GPU, encode/decode codecs, separate neural cores for ML tasks, shared memory, faster memory than the exchange from CPU to GPU on a PCIe x16 lane then back to CPU and back to memory. This all adds up. It is NOT just about the CPU portion and people need to realize this.

I can get the best Intel processor, but stick with standard SATA SSDs. Therefore, I will lose out of the 7 GB/s SSD speeds that Apple's ENTIRE CHIP can do. I can get the best Intel processor but stick an old GTX 1060 in there where Apple's GPU might beat that like crazy.

This. It's getting to the point where common benchmarks are increasingly misleading about actual performance. As you say, we are no longer comparing straight CPU performance, but different SoC configurations, some of which are optimized for specific tasks, or even software. Arguably, we are slowly moving away from general-purpose machines, at least if you want the best performance.

I would expect we will see some impressive results in Final Cut Pro for example, that don't align with raw benchmarks scores.

Apple has obviously cherry-picked their comparisons (e.g. 4x 5600M for Redshift), but overall performance should be pretty good. I look forward to the forthcoming YouTube reviews.

 

[huge_apple_fangirl]

This. It's getting to the point where common benchmarks are increasingly misleading about actual performance. As you say, we are no longer comparing straight CPU performance, but different SoC configurations, some of which are optimized for specific tasks, or even software. Arguably, we are slowly moving away from general-purpose machines, at least if you want the best performance.

I think this will be a huge advantage for Apple going forward. Improving IPC will get harder and harder in the future (and veeeery expensive) and accelerators will become even more important. Because Apple controls hardware and software, they will have a massive advantage. Look at the ProRes accelerator. Much harder to do that on Windows because you have to negotiate between many different parties. Apple's chips will be perfectly optimized for whatever real-world tasks they aim their machines at, irrespective of synthetic benchmarks (although I expect Apple to do well their too).

 

[Ethosik]

This. It's getting to the point where common benchmarks are increasingly misleading about actual performance. As you say, we are no longer comparing straight CPU performance, but different SoC configurations, some of which are optimized for specific tasks, or even software. Arguably, we are slowly moving away from general-purpose machines, at least if you want the best performance.

I would expect we will see some impressive results in Final Cut Pro for example, that don't align with raw benchmarks scores.

Apple has obviously cherry-picked their comparisons (e.g. 4x 5600M for Redshift), but overall performance should be pretty good. I look forward to the forthcoming YouTube reviews.

Yes, and I do think specialized SoC is the future instead of throwing more cores and threads on us. Otherwise in 10 years or so we all might be running 60 core systems.

I think Apple is doing things the right way here. The Final Cut Pro performance will be insane. As someone that uses Final Cut instead of Premiere (and I even pay for Premiere since I have all Adobe apps), this is exciting. I really do not like Premiere and really love Final Cut. I also love Logic and Garageband. So as long as Apple creates these SoC to handle these programs greatly, it will be all I need.

Also, I am still editing 1080p videos as 4K and even 8K does nothing for training videos in programming. In fact, it has a negative effect as a 1080p video blown up will have the larger syntax of programming vs native 4k or 8k where the syntax is very small. So even the cheapo M1 Mac mini is enough for me! I am hoping the new encoders/decoders will help even in my case as sometimes my recording sessions go for several hours so it takes several hours to export the video from Final Cut. So hopefully exporting is faster on these new chips.

 

[ultrakyo]

i just want to wait till they said. 1 nanometer chip .the last my friend said in taiwan max 2 or 1 not remember. now apple 5 and intel higher.Which mean intel hang long way.But if apple stuck those event would be troublesome of they make wafer bigger and bigger ...to support those event..hard moved actually.

1nm is already in the roadmap we just need to wait until it can be mass produced for Apple. What's after 1nm is still questionable but I am sure someone is already working towards a breakthrough. Maybe they would make the whole motherboard on a chip LoL.

TSMC And MIT Research Team Claims Amazing 1nm Chip Fab Breakthrough

 

[alien3dx]

1nm is already in the roadmap we just need to wait until it can be mass produced for Apple. What's after 1nm is still questionable but I am sure someone is already working towards a breakthrough. Maybe they would make the whole motherboard on a chip LoL.

TSMC And MIT Research Team Claims Amazing 1nm Chip Fab Breakthrough

yeah a bit scary .. then apple would said lighter and lighter ........ ?? .. what beyond 1 nanometer who know can be done or not.10 years ahead issue maybe.

 

[ian87w]

It potentially be yearly after completion of AS transition since now they control control the M series roadmap instead of relying on intel.

Apple religious event calendar should be like this:

March/April - Special Event (if any)
June - WWDC
September - A series Chip & Hardware
October - M Series Chip & Hardware

So we may see Mac Pro AS in March/April event at the earliest.

Imo Apple would/should consider the "upgrade cycle" of the Professional/Enterprise market as well. Although the consumer market is fine for having a yearly cycle, the Professional/Enterprise market might not be.

I have a feeling we will see a similar cadence like what's happening on the iPad. Consumer iPad gets yearly refreshes, while the iPad Pros have a longer refresh cycle. The Pro/Max Apple Silicon will still always win in GPU power compared to the regular M, even if they are behind in number. Just like the A chips, where the AnX versions have much more GPU power than the regular A version, allowing them to remain competitive with newer A chips.

 

[ian87w]

Raw CPU speeds are no longer a concern. It doesn't matter if Intel will have more cores or performance. People are still thinking these M1 and Pro/Max are simple CPUs. What they don't realize we have GPU, encode/decode codecs, separate neural cores for ML tasks, shared memory, faster memory than the exchange from CPU to GPU on a PCIe x16 lane then back to CPU and back to memory. This all adds up. It is NOT just about the CPU portion and people need to realize this.

I can get the best Intel processor, but stick with standard SATA SSDs. Therefore, I will lose out of the 7 GB/s SSD speeds that Apple's ENTIRE CHIP can do. I can get the best Intel processor but stick an old GTX 1060 in there where Apple's GPU might beat that like crazy.

Exactly. There are more to the story, and efficiency is the big thing in portables.

And Apple has basically shown how scalable their Apple Silicon is, being able to put the same M1 chip into the iPad and up to a desktop iMac, and using the same cores up to the M1 Max.

 

[ian87w]

Let’s see if they can get off 14nm first.

Tiger Lake is already 10nm. Pretty good for intel, and huge improvement over the 10th gen. But still a long stretch to match Apple Silicon efficiency.

 

[iPadified]

There is no evidence that there will be a two year cycle for Mac. There are only four basic chips to update each year: AX, MX, MX Pro and MX Max. MP will use a 2-4X MX Max. It is not as they start form the beginning with each chip. Mostly they are scaled versions of the AX with a few add-ons. It is very possible that Apple decide to rollout a new version every year.

 

[senttoschool]

Apple is concerned about profit, not about having the biggest market share. In the 1980s, they had multiple chances to capture most everything, but they chose to litigate, rather than innovate. In the 1990s, that previous strategy could have finished off the company. Ellen Hancock was probably the main reason Apple made it to all the good things that happened from 2000 onward.

Indeed, Apple is concerned about profits. Increasing market share (aka selling more products and services) is a great way to make more profits.

Reminds me of this:

And Steve is right. While there are $300 Dell systems, I would not recommend people buy them.

Macbook Air M1 was on sale for $850 on Amazon. Soon, we will have a Macbook SE for $750. It was not possible to have a good Macbook for a low price. Apple Silicon changed all that.

 

[G46&Fbnth5]

I don’t think M chips will get a clearly predictable upgrade cycle, unllike the base A series.

Looking at the past, we notice Apple didn’t make A7X, A11X and A13X, so I expect a similar approach to Macs (and iPad Pros): from time to time they’ll skip a generation to stay on pair with the technology used in iPhone chips.

 

[dumastudetto]

Apple has already won the war.

Intel are finished.

 

[alien3dx]

Apple has already won the war.

Intel are finished.

actually loosing . intel 10 nm while apple seem cool 5nm but each next year force to 4 3 ? 2 1 unsure unstable while 10 nm to 5 nm to 1nm give a lot of time for intel to optimize.

 

[leman]

actually loosing . intel 10 nm while apple seem cool 5nm but each next year force to 4 3 ? 2 1 unsure unstable while 10 nm to 5 nm to 1nm give a lot of time for intel to optimize.

Intel is currently effectively one node behind TSMC. And yet they need dour times more power to get the same performance as Apple gets out of their cores. Do you think that Intel's next node is going to be some sort of magic pixie dust that offers 4x power efficiency boost?

 

[quarkysg]

Intel is currently effectively one node behind TSMC. And yet they need dour times more power to get the same performance as Apple gets out of their cores. Do you think that Intel's next node is going to be some sort of magic pixie dust that offers 4x power efficiency boost?

No no no. I'm sure the latest and greatest Intel node processes will give you energy back

Sorry, couldn't help myself.

 

[Falhófnir]

Intel is currently effectively one node behind TSMC. And yet they need dour times more power to get the same performance as Apple gets out of their cores. Do you think that Intel's next node is going to be some sort of magic pixie dust that offers 4x power efficiency boost?

Wouldn't it be (almost) two given TSMC is now advancing beyond N5P? (and this is what M2 is likely to be built on) Intel is stuck on 10nm now (about 3 refinements in) while TSMC have gone down from a roughly equivalent 7nm/7nm+ to 5nm/5nm+ and are on the cusp of 4/3nm while Intel's true '7nm' is still nowhere in sight. While the exact correlation isn't there it seems TSMC are still advancing more rapidly.

If we look at transistor density:
Intel 10nm(++) = 106M/sq mm
TSMC 7nm(P) = 96-99M/sq mm
TSMC 5nm(P) = ~120-130M/sq mm
TSMC 4/3nm = (1.7x N5 = ~204M)