Geekbench score for the Intel i9-13900K is out and it’s not looking pretty for Apple …

[exoticSpice]

Hennessy believes that the future of hardware is domain-specific architecture, and Apple has a head start over AMD/Intel.

Where does he say that? What timestamp

 

[pshufd]

For sure, but Apple will still have the advantage of having the entire integration of hardware and software, and those optimizations cannot be underestimated.

If you are a content creator, those hardware encoders/decoders that are now in the base model M2 laptops are very nice. It's like you get an additional CPU that can run transcoding in the background without tying up your main CPU cores.

 

[Xiao_Xi]

Do they though? I mean, they have a slight head start in low-power inference hardware, but that's about it. One might mention Apple's AMX, but Intel has had AVX512 filling that niche for a while and they are releasing their own matrix accelerators (albeit for limited precision only) with the upcoming Xeon series. And I know from a little bird that AMD is working on integrating a vector processor, although not quite clear when it will happen. At any rate AMD has plenty of experience with vector stuff, being a GPU company and all.

I may be too influenced by Apple's marketing, but I believe that one of the advantages of Apple SoCs over AMD/Intel CPUs is purpose-built hardware accelerators, such as the neural engine and the media engine.

Last October, AMD said it would include more and more hardware accelerators in the coming years.

Where does he say that? What timestamp

I chose a bad wording. I meant to say that Hennessy believes that the future of computing is domain specific hardware (from 11:30 onwards), and I think Apple has a head start.

 

[pshufd]

I may be too influenced by Apple's marketing, but I believe that one of the advantages of Apple SoCs over AMD/Intel CPUs is purpose-built hardware accelerators, such as the neural engine and the media engine.

Last October, AMD said it would include more and more hardware accelerators in the coming years.

I chose a bad wording. I meant to say that Hennessy believes that the future of computing is domain specific hardware (from 11:30 onwards), and I think Apple has a head start.

The issue with Intel and AMD doing it is that they generally provide ISAs and I think that Apple is providing more like APIs. It seems to me like they are putting software into silicon instead of just giving you instruction sets that do specific, discrete pieces of work efficiently. What would be cool to me is a Huffman zig-zag decoder in hardware. Right now, it's very branchy code that could be done with vectors though it still requires a decent number of instructions. Why not just do a decode block in one instruction? The problem with putting in large amounts of special-purpose hardware is that these affect some users but not all. I'm sure that Oracle would love a lot of special purpose database silicon. But do you put that on all x86 chips that you sell?

Apple can probably get an idea of what people run on their systems and optimize for the programs that they believe require better performance. The creatives have to be pretty happy with the video accelerators in the M1 Pro, M1 Max and M1 Ultra. And now they come on the base model M2. How do you compete with giving away something that is really useful to a subset of people? In the old days, you bought the AfterBurner card for $5K.

 

[Unregistered 4U]

People might find this hard to accept, but Apple Silicon absolutely will be destroyed by Intel and AMD eventually. They're playing catchup but once they get to a 3 or 4nm things will change.

And, it won’t matter. Because, folks buy Macs because they want Macs. And they want the new one to be better than the one they currently own. There are very few people, if any, looking at their extensive investment in Apple peripherals, seeing that some random $1000+ chip scores better than Apple Silicon, and then buying a PC. So AMD and Intel could have raw benchmarks that are 2,3 or more times faster than Apple Silicon and, as NONE of those will be running the latest OS from Apple, they won’t matter.

The big hurdles the competition has are all power draw related, which is a solveable thing, and once solved will level the playing field and we'll be back to boring minor and fairly insignificant performance increases each year.

Oh, yes, it’s a solvable thing, BUT, the solution is not something they can actually do and remain viable in the market. In order to support ALLLLLLLL the Intel code written over the years, there’s a significant amount of work required for each x86 instruction just to decode it, to break it down to the point where the chip can do work on it. To solve the problem would be to simplify the instruction set, and to do that would require that a LOT of that old code breaks. If Intel makes the move, AMD would take marketshare, and vice versa if AMD decided to go first. So, they’re both locked in a battle of who can shovel crap faster.

 

[leman]

And, it won’t matter. Because, folks buy Macs because they want Macs. And they want the new one to be better than the one they currently own. There are very few people, if any, looking at their extensive investment in Apple peripherals, seeing that some random $1000+ chip scores better than Apple Silicon, and then buying a PC. So AMD and Intel could have raw benchmarks that are 2,3 or more times faster than Apple Silicon and, as NONE of those will be running the latest OS from Apple, they won’t matter.

I think you are way to optimistic. Sure, for some people it doesn't matter and they prefer the optics/haptics/long battery life of Apple computers. But many others? Macs are for example very popular in the developer community, exactly because they are such powerful and flexible tools that do not sacrifice ergonomy and battery for first class performance. So many devs choose Macs even though it can be an awkward platform for their products (which mainly revolves around Windows and Linux), and they build tools and environments to make Macs developer-friendly. But the moment x86 is 2x faster? Those people will abandon the Mac and won't look back. Heck, if x86 laptops were 2x faster even I would probably switch back to Linux, and I am a die hard Mac fan. But I also need to get work done and it does matter to me if my tools run 2x quicker.

 

[Unregistered 4U]

For feels you have PCIe 5.0 SSDs, not coming anytime soon to a Mac near you.

You mean these energy hungry heat-sink’d behemoths?

The first PCIe 5.0 SSDs for consumers are face-meltingly fast

Apacer reveals the first PCIe 5.0 SSD designs with 13,000MB/s write speeds and beefy heat sinks.

www.pcworld.com

Well, without a price or release date, doesn’t look like they’re coming ANYWHERE anytime soon

 

[Unregistered 4U]

They can easily afford to go up to 10-12W for their prosumer laptops and even higher for desktops. And they don’t care whether their chips are competitively priced, they can afford spending two or even three times more per chip than Intel does and they will still save money over buying x86 hardware on the market. Apple will always be able to afford more cache, more memory controllers, etc. etc.

But, unlike AMD and Intel, they actually care very much about their thermal performance as the thermal solution will be provided by Apple, not a customer. For every watt they allow, that’s one more watt they can NOT count on a customer buying an esoteric expensive cooling device for, Apple will be providing that solution.

 

[russell_314]

If your job is to run benchmarks a custom Intel PC is the way to go. You can pour N2 on the CPU and you'll get a crazy high score. This is really important if you're being paid on commission based on your scores. I can't seem to find that benchmarking job though.

 

[Unregistered 4U]

I may be too influenced by Apple's marketing, but I believe that one of the advantages of Apple SoCs over AMD/Intel CPUs is purpose-built hardware accelerators, such as the neural engine and the media engine.

I do, too. Apple could find out via analysis that there’s one process that’s in a lot of the code compiled for the app store that, if sped up, could be an improvement across the board, design the part for that process, and have it released in the next iteration. Neither AMD/Intel have that level of knowledge or flexibility. I’m expecting that whatever processor is in the AR/VR headset will have some custom hardware JUST for that use case that no one else will be able to shoehorn into their solutions for years. Raw benchmarks will show other solutions as faster, but apps coded for Apple’s glasses will still perform better than any other no-PC solution.

 

[leman]

But, unlike AMD and Intel, they actually care very much about their thermal performance as the thermal solution will be provided by Apple, not a customer. For every watt they allow, that’s one more watt they can NOT count on a customer buying an esoteric expensive cooling device for, Apple will be providing that solution.

I am not suggesting for Apple to go crazy like Intel does and draw 50 watts on a single core. But there would be nothing wrong with drawing up to 10 watts per core on a performance-oriented model. It would still be a fraction of what the competition uses and the chassis can absolutely handle it.

 

[Unregistered 4U]

I think you are way to optimistic. Sure, for some people it doesn't matter and they prefer the optics/haptics/long battery life of Apple computers. But many others? Macs are for example very popular in the developer community, exactly because they are such powerful and flexible tools that do not sacrifice ergonomy and battery for first class performance. So many devs choose Macs even though it can be an awkward platform for their products (which mainly revolves around Windows and Linux), and they build tools and environments to make Macs developer-friendly. But the moment x86 is 2x faster? Those people will abandon the Mac and won't look back. Heck, if x86 laptops were 2x faster even I would probably switch back to Linux, and I am a die hard Mac fan. But I also need to get work done and it does matter to me if my tools run 2x quicker.

It’s not optimism, it’s realism. There are a large number of developers, surely, but those numbers are dwarfed by the massive number of non-developers. At this point, they could all move to some other platform and it wouldn’t even show a blip… especially considering there would still be 34 million, more by now, developers using Macs to develop for Apple platforms. As long as the next system allows them to compile faster than the current one, they, too, would cast an uncaring eye towards whatever numbers Intel’s posting.

I should say it doesn’t matter to Apple or the vast majority of their users. It does matter to a subset that are using Macs for non-MacOS related purposes (like emulating other OS’s, developing non-Mac solutions, playing non-Mac games).

 

[Unregistered 4U]

I am not suggesting for Apple to go crazy like Intel does and draw 50 watts on a single core. But there would be nothing wrong with drawing up to 10 watts per core on a performance-oriented model. It would still be a fraction of what the competition uses and the chassis can absolutely handle it.

Oh, I completely agree they wouldn’t go to 50 watts, but every additional watt does bring more heat that has to be dealt with no matter how small it may seem. And, as long as the SoC group is in the same company as the software and the manufacturing groups, we’ll likely always see them trying to hit as low a temp as possible.

 

[leman]

It’s not optimism, it’s realism. There are a large number of developers, surely, but those numbers are dwarfed by the massive number of non-developers. At this point, they could all move to some other platform and it wouldn’t even show a blip… especially considering there would still be 34 million, more by now, developers using Macs to develop for Apple platforms. As long as the next system allows them to compile faster than the current one, they, too, would cast an uncaring eye towards whatever numbers Intel’s posting.

I should say it doesn’t matter to Apple or the vast majority of their users. It does matter to a subset that are using Macs for non-MacOS related purposes (like emulating other OS’s, developing non-Mac solutions, playing non-Mac games).

I don't think it's that simple. If the only people left on the platform are truly "lifestyle" users, then Apple is done as a computing company and becomes a gadget company. I mean, they would still make good money, but that will be the end of the Mac. Culture matters. And I think Apple agrees with me given the fact that their products aim at excellence in all the core areas and not just the least common denominator.

 

[leman]

Oh, I completely agree they wouldn’t go to 50 watts, but every additional watt does bring more heat that has to be dealt with no matter how small it may seem. And, as long as the SoC group is in the same company as the software and the manufacturing groups, we’ll likely always see them trying to hit as low a temp as possible.

A 16" MBP is capable of dissipating 80W of sustained heat with ease, it won't even turn the fans on before 30W or so. For such computer it doesn't matter whether a short-lived task draws 10W like now or 20W — the system can easily absorb it without any adverse effects. And it won't impact the battery either.

The fundamental problem with Intel CPUs is that they will burst into the stratosphere at the slightest provocation, briefly overloading the thermal system and kicking in the fans, no matter what you do. With its current per-core power consumption Apple has a lot of safe space to explore long before heat becomes an issue.

But hey, I'm just voicing my opinion. Nobody knows what they do. I'd just prefer them to offer higher peak performance for users who need it. They already have the best CPU tech in the world, it would be nice if it scaled a bit more to specific needs.

 

[pshufd]

I don't think it's that simple. If the only people left on the platform are truly "lifestyle" users, then Apple is done as a computing company and becomes a gadget company. I mean, they would still make good money, but that will be the end of the Mac. Culture matters. And I think Apple agrees with me given the fact that their products aim at excellence in all the core areas and not just the least common denominator.

Is a creative a lifestyle user? They used to require a decent amount of computing power but Apple's M1 put a floor on minimal performance. It's given a lot of people the ability to create YouTube videos and I'm sure that YouTube is seeing a lot of videos from it. I'd guess that this is happening in other areas as well.

Intel had this dog called the 10th generation and didn't get off their asses until M1 came along. Apple has an advantage for now and it appears that they will be pressing it.

Nobody noticed that R&D went from $5.7 billion to $6.8 billion YoY? That's a 19% increase in R&D. Sure, some of it was inflation but I doubt that all of it was. Intel spent $4.4 billion from $3.7 billion last year for an increase of 18%. So Intel is increasing R&D but I think that Apple has more cash flow to spend in that area.

 

[leman]

Is a creative a lifestyle user? They used to require a decent amount of computing power but Apple's M1 put a floor on minimal performance. It's given a lot of people the ability to create YouTube videos and I'm sure that YouTube is seeing a lot of videos from it. I'd guess that this is happening in other areas as well.

Great example. M1 has set an entirely new bar for minimal expectations. It's an entry-level (albeit premium) system that offers performance and features associated with high-end enthusiast PCs, without sacrificing the portability or usability. But that's exactly my point: it wouldn't be an attractive for context creators if it had bad performance. Sure, domain-specific accelerators do help, but you still need a decent enough GPU/CPU/memory etc. to do this kind of work. Remember that what I wrote was as reply to #304

Nobody noticed that R&D went from $5.7 billion to $6.8 billion YoY? That's a 19% increase in R&D. Sure, some of it was inflation but I doubt that all of it was. Intel spent $4.4 billion from $3.7 billion last year for an increase of 18%. So Intel is increasing R&D but I think that Apple has more cash flow to spend in that area.

Yep, and that's why I don't think that Intel will have an easy time catching up. Not to mention that Apple has the advantage of synergising their semiconductor efforts over their entire hardware ecosystem.

 

[Unregistered 4U]

But hey, I'm just voicing my opinion. Nobody knows what they do. I'd just prefer them to offer higher peak performance for users who need it. They already have the best CPU tech in the world, it would be nice if it scaled a bit more to specific needs.

Yeah, these days, they’re probably focusing more on the general user that makes up 90% of the folks that are buying their systems. If a few of the remaining folks want higher performance at the cost of not using macOS, I’m sure they’d welcome them doing so.

 

[pshufd]

Yeah, these days, they’re probably focusing more on the general user that makes up 90% of the folks that are buying their systems. If a few of the remaining folks want higher performance at the cost of not using macOS, I’m sure they’d welcome them doing so.

I think that Apple would love to have this market for bragging rights. But the volume market is where the profits are. Apple has the buzz right now and they are getting the word-of-mouth benefits from it. I would much rather see a 15 inch Air right now than a Mac Pro with an M2 Extreme chip. Or an M2 Pro Mac mini.

 

[Unregistered 4U]

I don't think it's that simple. If the only people left on the platform are truly "lifestyle" users, then Apple is done as a computing company and becomes a gadget company. I mean, they would still make good money, but that will be the end of the Mac. Culture matters. And I think Apple agrees with me given the fact that their products aim at excellence in all the core areas and not just the least common denominator.

It could be said that Apple was “done” as a computing company the day they removed “Computer” from their name. Look at those financials, what most people would call a “computer” doesn’t even make up a quarter of their revenue anymore! And when we consider unit sales, if one were to take a random sampling of folks that own Apple products (watches, iPhones, iPads), the vast majority of them wouldn’t own a Mac (and wouldn’t want to).

As long as there are people to buy Macs, but their importance to Apple centers around folks being able to develop code for iOS and running their pro software.

 

[Unregistered 4U]

I think that Apple would love to have this market for bragging rights. But the volume market is where the profits are. Apple has the buzz right now and they are getting the word-of-mouth benefits from it. I would much rather see a 15 inch Air right now than a Mac Pro with an M2 Extreme chip. Or an M2 Pro Mac mini.

Apple gave up bragging rights a long time ago. It was said that the war is over and Windows won, but admitting that doesn’t change the fact that Apple’s able to make a considerable amount of profit on something like 9% marketshare. Apple focuses on mobile because, like you, people in general value that mobility far more than any desktop specific features… sales across the board are 80% mobile, 20% desktop and are likely shifting in the direction of mobile every year.

 

[leman]

Yeah, these days, they’re probably focusing more on the general user that makes up 90% of the folks that are buying their systems. If a few of the remaining folks want higher performance at the cost of not using macOS, I’m sure they’d welcome them doing so.

I see no evidence for this. If that were true why bother with the prosumer chips, scalable interconnects, miniLED technology and thunderbolt? They are literally spending billions to cater to those users you say they don’t care about.

It could be said that Apple was “done” as a computing company the day they removed “Computer” from their name. Look at those financials, what most people would call a “computer” doesn’t even make up a quarter of their revenue anymore! And when we consider unit sales, if one were to take a random sampling of folks that own Apple products (watches, iPhones, iPads), the vast majority of them wouldn’t own a Mac (and wouldn’t want to).

It doesn’t matter how their revenue is structured, what matters is their culture and the brand perception. The Mac is central to all of that. And they are very much computing company at heart.

I mean, by that logic you must conclude that Microsoft doesn’t care about Windows since it’s just 15% of their revenue.

Apple gave up bragging rights a long time ago. It was said that the war is over and Windows won, but admitting that doesn’t change the fact that Apple’s able to make a considerable amount of profit on something like 9% marketshare. Apple focuses on mobile because, like you, people in general value that mobility far more than any desktop specific features… sales across the board are 80% mobile, 20% desktop and are likely shifting in the direction of mobile every year.

Apple never lost the PC war. They just had no interest to compete in the budget or server segment where 90% of all PCs are. Apple pretty much controls the premium PC market. You are confusing market share and presence. Ferrari or Lamborghini for example has zero market share on the car market. And yet they control desires and dreams of many car aficionado. Did you know that Ferrari makes almost as much money selling t-shirts and keychains as they do selling car?

As our old grandmaster Wong Shun Leung used to say: “look beyond the pointing finger”

 

[pshufd]

When the chips are down,

 

[jdb8167]

So many devs choose Macs even though it can be an awkward platform for their products (which mainly revolves around Windows and Linux), and they build tools and environments to make Macs developer-friendly.

Some good news today on that front. VMWare Fusion 22H2 Tech Preview was released with support for Windows 11. That along with solid Linux support, all for free for for personal, non-commercial use.

 

[pshufd]

These are rather stunning numbers. Even if these were all standalone businesses. Intel had $15 b in revenues.

Not bad for a Lifestyle company. I'm sure that Intel investors, for the past 15 years, wish that they made lifestyle products this good.