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

 

[mikegg]

M1X or M2X is going to massacre everything on the market. My goodness.

Buy Apple stocks. They're clearly poised to increase their laptop market share from 10%. It won't surprise me if they can reach 50% laptop market share in the U.S. in the next 5 years as they have done with the iPhone and iPad.

 

[Carfax83]

What people miss is that a wider design only matters when the code actually allows many instructions to be executed in parallel. Wider designs mostly shine during certain tight loops that can be unrolled/register renamed, and there are still diminishing returns the wider you get. Those tight loops also happen to the place where higher clock rates pretty much always work to your advantage.

I'm glad someone with actual knowledge said this. It seems like the novelty of such an exotic architecture as what is featured in the M1 has gone to the heads of many here on this forum. I guess we have to wait for more applications to become native to the M1 before we can start to see more meaningful comparisons.

But even so, for serious computing on a desktop x86-64 is going to be the only solution for a long time regardless. The M1's single core performance is awesome, but its multithreaded performance is wanting. And despite what the Apple fans say, multithreaded performance is just as, if not more relevant these days for desktop computing.....especially games.

Anyone thinking "oh Apple's SoCs perform as well as they do mainly because they are so wide" is missing all the other things they've done to reduce cache latency (though obviously, going lower when measured in cycles is easier when your cycle time is 1/3.2 GHz instead of 1/5.0 GHz) improve branch prediction (apparently better than Intel's, which had been considered the gold standard is prediction efficiency) and countless other little things that alone probably aren't even close to a 1% improvement but collectively add up.

One thing I have to applaud Apple for, is their integration. It's masterful indeed, and I'm sure it adds to their overall efficiency and performance.

 

[Carfax83]

M1X or M2X is going to massacre everything on the market. My goodness.

How are they going to massacre everything when their multithreaded performance is so far behind?

 

[mikegg]

How are they going to massacre everything when their multithreaded performance is so far behind?

1. Because of M1X/M2X. Maybe one day a Mac Pro level SoC as well but that's a niche product. M1 is/will be Apple's the slowest laptop SoC.
2. Multicore is already competitive with 4900HS, according to Anandtech's review.
3. While Intel and AMD's chips have to brute force everything, Apple can uses hardware acceleration for traditionally multithreaded apps, leaving Intel and AMD in the dust. Scroll up 3 posts and see the benchmarks of real-world applications I posted.
4. Apple can actually get developers to make use of hardware accelerators because they control both the hardware and software.
5. In a pure 1-core to 1-core comparison, only the AMD 5950x is close to Apple's entry-level M1, according to Anandtech.
6. In everyday tasks such as opening new apps or checking email or web browsing, the M1 is far more responsive than Intel's very best mobile chips (and by extension, AMD).
7. Most computers sold are laptops and Apple Silicon offers a significantly better product than AMD/Intel here.
8. Unlike people on Anandtech forums, people don't care about whether things are done on the CPU or GPU or a dedicated accelerator in the SoC. If it's faster, it's faster.

 

[insertcarehere]

But even so, for serious computing on a desktop x86-64 is going to be the only solution for a long time regardless. The M1's single core performance is awesome, but its multithreaded performance is wanting.

Compared to what? The M1 handles itself just fine in MT workloads comparing to other chips that fit in a similar form factor. Penalizing the M1 for not keeping up with a workstation desktop with 10x the footprint is silly, assuming the M1 represents a "upper bound" for Apple's overall MT performance equally so.

And despite what the Apple fans say, multithreaded performance is just as, if not more relevant these days for desktop computing.....especially games.

Hardware acceleration is also just as relavant nowadays given how many real world MacOS apps seem to benefit from it. If we dock points for the M1's MT performance in some workloads then for consistency the M1's performance when taking advantage of hardware acceleration in other workloads should not be dismissed either.

 

[Roland00Address]

This is kind of dancing around the terminology if you ask me. If the smaller more efficient cores are being tapped during multithreaded workloads, then the M1 is an 8 core CPU no matter which way you slice it.

But Apple has marketed the M1 as having the "World's fastest CPU core" so they should be able to back that up if you ask me:

Since you are taking this “literally“ and you feel offended for you feel you are being lied to...

I should point out Apple said the world’s fastest cpu core as in singular core and not cores, soc, chip, etc. Thus multi core performance is not apple’s claim.

 

[Carfax83]

1. Because of M1X/M2X. Maybe one day a Mac Pro level SoC as well but that's a niche product. M1 is/will be Apple's the slowest laptop SoC.
2. Multicore is already competitive with 4900HS, according to Anandtech's review.
3. While Intel and AMD's chips have to brute force everything, Apple can uses hardware acceleration for traditionally multithreaded apps, leaving Intel and AMD in the dust. Scroll up 3 posts and see the benchmarks of real-world applications I posted.
4. Apple can actually get developers to make use of hardware accelerators because they control both the hardware and software.
5. In a pure 1-core to 1-core comparison, only the AMD 5950x is close to Apple's entry-level M1, according to Anandtech.
6. In everyday tasks such as opening new apps or checking email or web browsing, the M1 is far more responsive than Intel's very best mobile chips (and by extension, AMD).
7. Most computers sold are laptops and Apple Silicon offers a significantly better product than AMD/Intel here.
8. Unlike people on Anandtech forums, people don't care about whether things are done on the CPU or GPU or a dedicated accelerator in the SoC. If it's faster, it's faster.

1) So you're alluding that you already know the specs of the M1X/M2X and what it will be called?
2) Zen 2 is over a year old at this point. If or when the M1X as you put it comes out, it won't be competing against Zen 2. It will be a Zen 3/Golden Cove variant or maybe even a Zen 4.
3) So you think that hardware acceleration doesn't exist on PC? LOL! The benchmarks that you posted above show just how much more powerful a desktop GPU can be compared to an integrated one.
4) Hardware acceleration has been around on PC for a long time now. It works very well for certain things.
5) Yes, that is definitely the M1's strength; raw single threaded performance. But single core performance is much less relevant on desktop/workstation/server platforms.
6) I don't buy this at all. Even on my old PC, everything is snappy and web browsing is practically instantaneous due in part to utilizing my GPU. Browser benchmarks are a joke, unless you're on mobile.
7) I won't deny that the M1 is a great product, but many (including myself) are turned off by Apple's walled garden and will never buy it no matter how good it is.
8) Again, you seem to think that hardware acceleration is an Apple thing only, when it has been around on PC for a long time now.

Your own benchmarks that you posted above shows how much of an impact hardware acceleration can have, even on an aging product like the Intel Mac mini.

 

[Carfax83]

Compared to what? The M1 handles itself just fine in MT workloads comparing to other chips that fit in a similar form factor. Penalizing the M1 for not keeping up with a workstation desktop with 10x the footprint is silly, assuming the M1 represents a "upper bound" for Apple's overall MT performance equally so.

If you look at the context of that comment I made, I was obviously referring to high performance desktops. Assuming Apple ever wants to design ARM based high end Macs, they will have to up their game in the multicore performance.

After the M1, I have no doubt that Apple could design an APU with 8 big cores that would scale better.

Hardware acceleration is also just as relavant nowadays given how many real world MacOS apps seem to benefit from it. If we dock points for the M1's MT performance in some workloads then for consistency the M1's performance when taking advantage of hardware acceleration in other workloads should not be dismissed either.

You won't get any disagreement from me. Hardware acceleration is definitely relevant and very useful. But as I told @senttoschool earlier, hardware acceleration has been around on PC for a very long time now so it's nothing new. In fact, the hardware acceleration available on PC is much more powerful than what you get in the M1.

 

[beginner99]

It won't surprise me if they can reach 50% laptop market share in the U.S. in the next 5 years as they have done with the iPhone and iPad.

I very much doubt that because laptops aren't a status symbol like smartphones and hence people mostly go for the cheap crap also because there aren't any "provider scams" to get the phone for free while paying ridiculous monthly fees for several years.

 

[Carfax83]

Since you are taking this “literally“ and you feel offended for you feel you are being lied to...

Why would you think I feel offended? I was just stating what I think is factual. The M1 is an 8 core CPU.

Are we going to say that Intel's Golden Cove with 8 big cores and 8 little cores is not a 16 core CPU as well? If it's doing work, it should be counted!

I should point out Apple said the world’s fastest cpu core as in singular core and not cores, soc, chip, etc. Thus multi core performance is not apple’s claim.

Good point.....but that is still contested as Zen 3 obviously can match it in single threaded performance.

 

[beginner99]

f you look at the recent benchmarks that @senttoschool posted above, you can see why hardware acceleration can make doing valid architectural comparisons futile.

It's not just the architecture but the whole platform from SOC to drivers/API to the OS. Apple controls it all and there is a very limited number of configs. Its essentially like it used to be with consoles. If you control everything you can extract far more from the hardware especially dedicaded hardware that gets called by the APIs automatically without developer having to do anything special. Bascially AMDs fusion dream that never really came to fruition, instead of AVX the gpu could be used transparently for matrix calculations. But that simply doesn't work in the x86 world.

 

[mikegg]

I very much doubt that because laptops aren't a status symbol like smartphones and hence people mostly go for the cheap crap also because there aren't any "provider scams" to get the phone for free while paying ridiculous monthly fees for several years.

No one thinks phones are status symbols anymore in the US. Maybe in a 3rd world country. You can buy an iPhone for $400.

 

[fobis]

1. Because of M1X/M2X. Maybe one day a Mac Pro level SoC as well but that's a niche product. M1 is/will be Apple's the slowest laptop SoC.
2. Multicore is already competitive with 4900HS, according to Anandtech's review.
3. While Intel and AMD's chips have to brute force everything, Apple can uses hardware acceleration for traditionally multithreaded apps, leaving Intel and AMD in the dust. Scroll up 3 posts and see the benchmarks of real-world applications I posted.
4. Apple can actually get developers to make use of hardware accelerators because they control both the hardware and software.
5. In a pure 1-core to 1-core comparison, only the AMD 5950x is close to Apple's entry-level M1, according to Anandtech.
6. In everyday tasks such as opening new apps or checking email or web browsing, the M1 is far more responsive than Intel's very best mobile chips (and by extension, AMD).
7. Most computers sold are laptops and Apple Silicon offers a significantly better product than AMD/Intel here.
8. Unlike people on Anandtech forums, people don't care about whether things are done on the CPU or GPU or a dedicated accelerator in the SoC. If it's faster, it's faster.

I perfectly understand the desire to compare the performance of Apple Silicon versus Intel and AMD offerings. There are a lot of Interesting facets to consider: x86 vs ARM (CISC vs RISC, by extension), IDM vs foundry, general CPU design, manufacturing node, etc. Lots of topics on which people can be armchair experts, happily clacking away at their keyboards (present company included).

What I don’t understand is why there are now ~50 pages of posts pretending that Apple’s foray into PC chips even matters to most potential customers - that Intel and AMD are now doomed. Apple has long held a lead in performance on their A-series chips. Enthusiasts often lament the lackluster offerings of Qualcomm when compared to Apple’s chips, yet continue to happily purchase their Android-based smartphones. Why isn’t Apple’s worldwide marketshare for smartphones near-monopolistic when they maintain a clear lead in SoC design and performance? I would argue that it is for the same reason their worldwide marketshare for laptops nor desktops will not skyrocket with the new M-series chips: the CPU/SoC is not important as the entirety of the widget they power, the ecosystem in which that widget exists, and the ecosystem which that widget provides.

Apple’s advances are significant. The industry is paying attention. Enthusiasts are paying attention. But the hyperbole about how Apple will henceforth dominate Intel and AMD is myopic. Do the Apple chips perform admirably in real usage and benchmarks? Absolutely. Do they beat Intel and AMD offerings in certain scenarios? You bet. Will many data centers suddenly switch all of their systems to stacks of M1 Mac minis? Probably not. Will a bunch of businesses dump their fleet of Windows systems for M1 MacBooks? Doubtful. Will individual customers shopping in the sub-$600 laptop market decide that the M1 is worth the extra cash? No.

Apple Silicon will continue to live in a bubble where comparisons to chips from Intel, AMD, Qualcomm, Samsung, MediaTek, HiSilicon, etc., will remain largely academic because Apple doesn’t directly compete with those companies - Apple competes with other widget-makers who use chips from those other companies. Those other companies don’t have to worry about competition from Apple. They need to worry about where Apple is signaling the market to go: more control of the entirety of the widget to differentiate products from the competition. They need to worry about their customers (who have the money and ability) that could start designing their own specialized chips for their own widgets.

Edit: D’oh. Samsung is obviously a competitor to Apple. HiSilicon is owned by Huawei, who is also a competitor to Apple.

 

[beginner99]

No one thinks phones are status symbols anymore. You can buy an iPhone for $400.

See they got you already because you think $400 is cheap when 5 years ago the flagships cost $500 (which are now over $1000 up to $1500). I have to admit however that in the high end market apple actually makes more sense because they certainly have better performance, apple certainly has the "luxury brand" feel to it.

Besides that paying $400 for a phone with huge bezels and tiny screen, well it's subjective. Certainly a no-go for me.

(smartphone pricing is just as crazy as gpu where people now say $400 is midrange when 5 years ago $500 got you the flagship.)

 

[Qwertilot]

I'm glad someone with actual knowledge said this. It seems like the novelty of such an exotic architecture as what is featured in the M1 has gone to the heads of many here on this forum. I guess we have to wait for more applications to become native to the M1 before we can start to see more meaningful comparisons.

But even so, for serious computing on a desktop x86-64 is going to be the only solution for a long time regardless.

Massacre is rather emotive for the 'bigger' M1 variant of course but so is the idea that you couldn't use M1 for serious computing The 16gb is the latest built in limitation.

The bigger chip is still going into laptops so will have some limits, especially as they'll want to keep it integrated and relatively sane power draw.

So, 2-3 times the power budget of M1 to play with.

Quite likely quad core ram to let it go up to 32gb.

They could plausibly double MT cpu performance over M1, and up ST a little bit. Similar for the already rather impressive integrated gpu to 'compensate' for the dGpu's.

Not sure about the npu but if they up that a chunk and it has up to 32gb of memory to use then it could be awfully fast at some ML tasks.

As a laptop chip it'll be broadly terrifying. Still, it won't be a really top end workstation chip. That will have to wait a while.

I definitely wouldn't take comfort in YoY improvement by other people though - Apple have been producing very regular, large increases for years now.

There's no reason to expect that to stop. It won't feed immediately into Mac chips, maybe every 2 years, but it will go in.

 

[jeanlain]

But Apple has marketed the M1 as having the "World's fastest CPU core" so they should be able to back that up if you ask me:

They claim the M1 to be the fastest CPU core, when it comes to low-power Silicon.

Mac

The most powerful Mac laptops and desktops ever. Supercharged by Apple silicon. MacBook Air, MacBook Pro, iMac, Mac mini, Mac Studio, and Mac Pro.

www.apple.com

1. Testing conducted by Apple in October 2020 using preproduction 13‑inch MacBook Pro systems with Apple M1 chip and 16GB of RAM measuring peak single-thread performance of workloads taken from select industry-standard benchmarks, commercial applications, and open source applications. Comparison made against the highest-performing CPUs for notebooks commercially available at the time of testing. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

Which production laptop CPU had a faster core than the M1 in October? Even today, in SPEC tests, the M1 is clearly faster than Tiger Lake, same in geekbench (of course). I think the M1 core beats the intel core more often than not.

 

[jeanlain]

It's not just the architecture but the whole platform from SOC to drivers/API to the OS. Apple controls it all and there is a very limited number of configs. Its essentially like it used to be with consoles. If you control everything you can extract far more from the hardware especially dedicaded hardware that gets called by the APIs automatically without developer having to do anything special. Bascially AMDs fusion dream that never really came to fruition, instead of AVX the gpu could be used transparently for matrix calculations. But that simply doesn't work in the x86 world.

Apple also heavily relies on high level frameworks to abstract away the hardware (they're not the only ones of course). For instance, Apple has the "Accelerate" framework for common mathematical calculations using vectors and matrices. Accelerate traditionally uses the SIMD units on the CPU, regardless of its type (PPC, Intel, M1). It can even make use of the neural engine or other specialised hardware. Developers don't even have to know about it.
The same is true for VideoToolBox. It uses whatever accelerator is available to decode/encode video. You don't need to worry about the presence of Apple/Intel/AMD hardware.

 

[mikegg]

Apple Silicon will continue to live in a bubble where comparisons to chips from Intel, AMD, Qualcomm, Samsung, MediaTek, HiSilicon, etc., will remain largely academic because Apple doesn’t directly compete with those companies - Apple competes with other widget-makers who use chips from those other companies. Those other companies don’t have to worry about competition from Apple. They need to worry about where Apple is signaling the market to go: more control of the entirety of the widget to differentiate products from the competition. They need to worry about their customers (who have the money and ability) that could start designing their own specialized chips for their own widgets.

This is wrong. Apple absolutely competes against Intel, AMD, Qualcomm, Samsung, MediaTek, HiSilicon, etc. In addition, Apple switching to ARM has huge implications.

1. It allows Apple to decrease prices, ala iPhone SE/Watch SE, and reach out to the budget crowd, and still offer a better experience than budget Windows PCs. Apple owns 50% of the phone market in the U.S., 65% of the tablet market, and most of the Smart Watch market. There is no reason why normal people (not PC master race Anandtechers) won't buy Macs if they're significantly better.

2. It allows the ARM software ecosystem to move beyond mobile. It's very attractive to developers if they can develop one piece of software and have it work on laptops/desktops/phones/tablets. Expect Microsoft ARM to get more software support. Expect Linux to better support ARM. An example is Docker supporting ARM purely because of Apple Silicon.

3. When server software is more compatible with ARM, expect Amazon, Google, Microsoft, etc to start offering more ARM servers. AWS' Graviton2 is already competitive or better than Epyc in many types of loads when you measure cost/performance.

This is Anandtech's own conclusion as well:

For developers, the Apple Silicon Macs also represent the very first full-fledged Arm machines on the market that have few-to-no compromises. This is a massive boost not just for Apple, but for the larger Arm ecosystem and the growing Arm cloud-computing business.

 

[mikegg]

See they got you already because you think $400 is cheap when 5 years ago the flagships cost $500 (which are now over $1000 up to $1500). I have to admit however that in the high end market apple actually makes more sense because they certainly have better performance, apple certainly has the "luxury brand" feel to it.

Besides that paying $400 for a phone with huge bezels and tiny screen, well it's subjective. Certainly a no-go for me.

(smartphone pricing is just as crazy as gpu where people now say $400 is midrange when 5 years ago $500 got you the flagship.)

Every major phone maker has phones costing $1000 nowadays. It's not just Apple.

No one sees phones as a status symbol anymore. Neither does Apple.

By the way, the iPhone SE at $400 is faster than any Android phone you can buy.

 

[fobis]

This is wrong. Apple absolutely competes against Intel, AMD, Qualcomm, Samsung, MediaTek, HiSilicon, etc. In addition, Apple switching to ARM has huge implications.

1. It allows Apple to decrease prices, ala iPhone SE/Watch SE, and reach out to the budget crowd, and still offer a better experience than budget Windows PCs. Apple owns 50% of the phone market in the U.S., 65% of the tablet market, and most of the Smart Watch market. There is no reason why normal people (not PC master race Anandtechers) won't buy Macs if they're significantly better.

2. It allows the ARM software ecosystem to move beyond mobile. It's very attractive to developers if they can develop one piece of software and have it work on laptops/desktops/phones/tablets. Expect Microsoft ARM to get more software support. Expect Linux to better support ARM. An example is Docker supporting ARM purely because of Apple Silicon.

3. When server software is more compatible with ARM, expect Amazon, Google, Microsoft, etc to start offering more ARM servers. AWS' Graviton2 is already competitive or better than Epyc in many types of loads when you measure cost/performance.

This is Anandtech's own conclusion as well:

For developers, the Apple Silicon Macs also represent the very first full-fledged Arm machines on the market that have few-to-no compromises. This is a massive boost not just for Apple, but for the larger Arm ecosystem and the growing Arm cloud-computing business.

1. Apple choosing to serve a lower cost segment of the market with certain products does not make them a competitor with all other chip designers. It makes them a competitor with other widget makers in that market segment. “Normal people” make purchasing decisions based on a myriad of rational and irrational reasons. “Significantly better” is ambiguous. If “normal people” always purchased “significantly better” products, then there would be zero vehicles sold with ICEs today. But wait - could you argue that electric vehicles aren’t “significantly better” than their gas-guzzling brethren? You sure could in a certain view of reality. As an ardent Apple fan who uses an iPhone and iPad as their main computing devices, I promise you that Apple devices are not “significantly better” to everyone.

2 and 3. You are adding to my argument that chip designers need to worry about losing customers from the industry shifting toward in-house designs. Apple has shown that there is potential value in breaking away from the traditional chip designs and designers, tailoring a chip to their particular needs as opposed to choosing from a sea of SKUs. As you mentioned, the big tech companies have already started designing their own chips (many based on ARM but also ASICs) in earnest. Much of this pre-dated M1, but I would argue that Apple’s A-series chips were at the forefront of this movement for general application processors.

You have not adequately refuted my assertions. I will concede that I incorrectly listed HiSilicon and Samsung (LSI) as not competitors to Apple. HiSilicon is wholly owned by Huawei, who is a competitor to Apple. Samsung is obviously a competitor to Apple.

 

[moinmoin]

Looks like M1's already limited I/O has some teething problems, though it's unclear so far how much of it is due to hardware or software issues:

tldw:

• frame drops and distorted audio in Final Cut Pro (reboot seems to fix it)
• USB slower than on Intel Macs (read speed measured: 780 MB/s on MBP M1 compared to 910 MB/s on iMac i5)
• USB speed doesn't seem to be constant (used ethernet on a dock for file transfer, speed drops out regularly)
• external display up to 60 Hz only on HDMI
• external display on Thunderbold or DP only supports up to 30 Hz
• when connected internal display can turn off without a way to recover (reboot seems to fix it)

 

[jeanlain]

Looks like M1's already limited I/O has some teething problems

Yes, early software can have bugs. It's way too early to conclude that these issues are due to supposed M1 limitations regarding I/O.
Regarding FCPX, if the M1 was limited in its throughput, it would be seen accross the board, not just in certain scenarios. Lots of people already use the M1 to edit their videos in 4k, and the overall message is that the M1 is excellent. As mentioned in the comments, there could be some issues with the particular sony codec used by this youtuber.
As for the 60Hz only via HDMI... The M1 can drive a 6K display at 60Hz via thunderbolt. I'm not sure why this guy has a problem with his monitor. I'm sure other monitors work fine at 60Hz without using HDMI. It can't be a limitation of the M1 per se.
As for USB speed... I don't know. This could also be related to driver issues. I'm sceptical that the USB-4 ports are somehow slower than the UBS-C/TB3 ports of the previous models.

 

[beginner99]

By the way, the iPhone SE at $400 is faster than any Android phone you can buy.

It's not the only metric that matters in a smartphone. only reason I replaced my 5 year old phone was because I fell on it and the screen broke and repraing obviously isn't worth it at that point. It was fast enough to browse the net even if it probably was 5x times slower than an M1 in geekbench and the likes.
I paid less for my new phone than $400, it has 5g (somewhat relevant because I will keep it as long as possible) and more storage, better screen and screen resolution and much better design (subjective). Plus the iPhone SE is more like $450(65gb)-$550(128gb) here so I paid more than $150 less for comparable features at cost of slower performance.

 

[moinmoin]

Yes, early software can have bugs. It's way too early to conclude that these issues are due to supposed M1 limitations regarding I/O.

Why do you cut the quoted sentence in half to proceed to state the same?

 

[Leeea]

Lots of people already use the M1 to edit their videos in 4k, and the overall message is that the M1 is excellent.

I feel for the poor sap who has to edit a 4k video with 16 GB of ram.

...

I think it was Linus Tech Tips complaining that Threadripper could only take 256 GB of ram, and that was just barely enough for his 4k video editing.