Ah yes, breaking news - coffee shop owner predicts that coffee will overtake water as the most consumed beverage in the world by 20XX. “Chances are good,” he says, “especially considering we have an exclusive agreement with McDonalds to pour coffee into customers cup regardless of their order”.
I am obviously being extremely snarky here, but such predictions are usually based on providing the best optics for shareholders.

Not sure... but what keeps Intel from making ARM too? They actually do already in Alteras... so what's the problem?

From where these Yellow press ideas come from?

Ferrum MasterNot sure... but what keeps Intel from making ARM too? They actually do already in Alteras... so what's the problem?

From where these Yellow press ideas come from?

If they licensed ARM cores, they'd have to pay royalties, giving up much of their profit to Arm - so losing to Arm in a sense.
They could however adopt the RISC-V arch, and fight Arm that way.

but yeah, this reads like shameless marketing/shareholder courting

At the moment Arm has an 11% share of the Windows PC market according to Counterpoint Research.

That is also an interesting point in the source article.

I think it’s far more likely Qualcomm and mediatek will be moving to risc v for a good chunk of their products and arms market share dropping significantly overall that it is them taking over the pc market

londisteThat is also an interesting point in the source article.

maybe it’s Microsoft surface laptops being sent to vendors that day?

PastravThey could however adopt the RISC-V arch

RISC-V is definitely more powerful and energy efficient than the CISC x86-64.
It would be really good for everyone if the CISC arch is abandoned and forgotten once and for all.

Qualcomm seems to have all the support from Microsoft and big OEMs like Dell. And seeing Dell supporting Qualcomm is too alarming for x86 and Intel particularly. I guess Dell is trying to secure itself in case Intel's fab business fails and with it, Intel fails. Also no wonder Dell keeps downplaying the existence of AMD and goes with Qualcomm today, probably Nvidia tomorrow. Dell was always favoring the bigger company as their (almost exclusive) partner.

And we haven't seen Nvidia yet entering the game. Nvidia could use it's power in GPUs to push the ARM platform to gamers and enthusiasts. Imagine, for example, RTX 7000 being faster or cheaper or both in an "optimized for RTX" ARM based platform from Nvidia. Many will be swapping to ARM to get all the performance of their highly performing and slightly less ridiculously expensive RTX 7090 for ARM, over the RTX 7090 for x86. I was saying for years that Nvidia is the company that could move people to ARM over x86. And Nvidia's position today in the gaming GPU market, combined with AMD's lack to be highly competitive in the hi end gaming market, does offer Nvidia the opportunity to become a company that sells the whole platform to the consumer, not just the GPU. I doubt Nvidia doesn't consider this.

Then we have Mediatek that can cover the cheaper models. Laptops that offer over 20 hours on battery with prices closer to $600 that the $1000 for the Qualcomm models.

There is huge support behind ARM this time and Apple already proved that moving from x86 to ARM, is not impossible.

Not to forget that the laptop market is much bigger than the desktop market and maybe that 50% by 2029 is not plain wishful thinking or marketing BS, but something that could be archived.


P.S. And of course no one stops AMD from making ARM SOCs and APUs.

I think it's getting to a point where x86 is no longer strictly necessary to create a reliable, high-performance processor, and with Windows desperately needing an exit from this backwards compatibility baggage that plagues it, the architecture has become cumbersome in itself.

Intel is likely aware of this, which is why they've come up with x86-S proposal (I do not think that will stick). Licensing x86 out might be Intel's salvation here, if they mean to keep this architecture around. However, it might be in Intel's interests to design ultra high performance ARM ISA-compatible cores in the future, somewhat like Apple Silicon but on steroids.

All in all, x86-64 CPUs ain't going anywhere, at least not for another 10 years or so. 2029 is boastful talk from the CEO at best, and delusional at worst, and would assume that ARM chips are already widespread in the PC landscape today, which couldn't be further from the truth. The usage of ARM in Mac is consequential to Apple's choice to unify hardware to strengthen their vertical integration and reduce reliance on third-party companies for their components, but for PC, the Snapdragon X Elite is still a no-show and the 7cx processors are found strictly in the ultra low cost PC segment (formerly netbooks) in some of the cheapest Compaq Presarios you can get today.

john_P.S. And of course no one stops AMD from making ARM SOCs and APUs.

In the past, AMD's Opteron ARM processors were some of the most commercially unsuccessful products the company has ever launched. Would be interesting how that would pan out today, though.

Ferrum MasterNot sure... but what keeps Intel from making ARM too? They actually do already in Alteras... so what's the problem?

It's in their best interest x86 to remain the platform of choice. So I am thinking they will jump to ARM last. But they are willing to build ARM chips in their fabs, so if their fab business becomes their core business, they probably wouldn't mind losing that x86 advantage they have.

londisteThat is also an interesting point in the source article.

I wonder how they came up to that number. I mean, if ARM based laptops where half the market share of AMD's in laptops, we would easily spot a few laptops here and there. Where are those?

john_Also no wonder Dell keeps downplaying the existence of AMD

john_Dell was always favoring the bigger company as their (almost exclusive) partner.

AMD is a bigger and more stable company than intel. And even Qualcomm.

AMD:


Qualcomm:


intel:

john_I guess Dell is trying to secure itself in case Intel's fab business fails and with it, Intel fails.

Err, they can always continue to make 22 and 32nm products for lower performance devices.

ARFAMD is a bigger and more stable company than intel. And even Qualcomm.

AMD:


Qualcomm:


intel:




Err, they can always continue to make 22 and 32nm products for lower performance devices.

If regulatory agencies weren't a thing, JHH could buy all 3 with candy money if we were to bring a company's valuation into question. $NVDA is currently valued at 3T, after all. I don't think that stock prices are an indication of a company's stability, just that its products are currently successful, in vogue and in the interests of investors.

Intel specifically has undervalued shares, because the company has extreme CapEx and has committed some blunders recently that have affected investor view, at least in the midst of the segment that understands the company's products. If I had to pick any of the 4 big tech stocks, the fourth being $NVDA, the one I'd buy right now is $INTC because it's in a low and it has the biggest growth potential behind NVIDIA (though $NVDA will be purchasable again after the 10-1 split occurs, can't spend $1280 in a single paper).

Dr. DroIn the past, AMD's Opteron ARM processors were some of the most commercially unsuccessful products the company has ever launched. Would be interesting how that would pan out today, though.

It was the Bulldozer era of AMD, AMD probably didn't had money and engineers to spare for something meaningful. I mean, I doubt they where seriously playing with customizing those standard ARM cores.

john_It was the Bulldozer era of AMD, AMD probably didn't had money and engineers to spare for something meaningful. I mean, I doubt they where seriously playing with customizing those standard ARM cores.

IIRC they went with Cortex-A57 cores back then. No clue if they were custom or not

ARFAMD is a bigger and more stable company than intel. And even Qualcomm.

Meaningless numbers.
Intel's bonds with Dell are known for over 20 years and Intel having fabs means that it can secure a huge number of CPUs and chipsets that Dell needs. AMD can't do that.
Qualcomm is the stronger company in the ARM platform - ignoring Apple here - and also with too strong leverage in that platform, thanks also to their superior networking hardware. One of the reason's Nvidia stopped making SOCs for smartphones about 10 years ago - and this is very very funny - is because Qualcomm was using it's power and anticompetitive tactics against it's competitors, with Nvidia being one of those competitors.

Err, they can always continue to make 22 and 32nm products for lower performance devices.

You can ask GlobalFoundries how good they are doing after dropping their plans for more advanced nodes, especially now that China is turning itself in the biggest... forest of fabs for older nodes.

ARFRISC-V is definitely more powerful and energy efficient than the CISC x86-64.
It would be really good for everyone if the CISC arch is abandoned and forgotten once and for all.

CISC vs RISC argument has been essentially dead since micro-ops became a widespread thing.
The success RISC-V has had in adoption has nothing to do with anything related to ISA or architecture. RISC-V is free. This is their winning argument :)

Dr. DroIIRC they went with Cortex-A57 cores back then. No clue if they were custom or not

I think they where not custom cores. And even if they where, I doubt AMD was throwing money and engineers on that project. Probably they just wanted to be somewhat ready in case ARM was becoming a success in laptops, desktops and/or servers. Su killing that project when she took over, probably means that they weren't looking at it seriously. So it was just a money drain that needed to stop.

Dr. Dro$INTC because it's in a low and it has the biggest growth potential behind NVIDIA

I can see Intel at $100 if their fab business ends up competitive to TSMCs. But they will have to also move fast to ARM if they see x86 losing market share and companies like Microsoft moving away from it.

john_I think they where not custom cores. And even if they where, I doubt AMD was throwing money and engineers on that project. Probably they just wanted to be somewhat ready in case ARM was becoming a success in laptops, desktops and/or servers. Su killing that project when she took over, probably means that they weren't looking at it seriously. So it was just a money drain that needed to stop.

They were not. They were very basic 4 or 8 A57 configs with two 10Gbe NICs and, like, 8 PCIe lanes. They existed seemingly solely as a financial lifeline.
Now, whether canceling the initiative altogether was a wise move is debatable. I remember couple of years ago Jim Keller explicitly called out AMD not investing into ARM server chips as short-sighted, which is an understandable point of view seeing where the industry is going.

OnasiThey were not. They were very basic 4 or 8 A57 configs with two 10Gbe NICs and, like, 8 PCIe lanes. They existed seemingly solely as a financial lifeline.
Now, whether canceling the initiative altogether was a wise move is debatable. I remember couple of years ago Jim Keller explicitly called out AMD not investing into ARM server chips as short-sighted, which is an understandable point of view seeing where the industry is going.

I think it was the right choice. AMD would had needed to keep throwing money for the last 10+ years only to see ARM (PROBABLY) becoming a real competitor to x86 in servers and laptops this year for the first time. If AMD starts investing today, money that they DO have today, they might be ready in 3-5 years. Until then Ryzen and EPYC will keep selling.
Keller's prediction was like the prediction from 2000 that Linux will beat Windows. Way too early.

ARM CEO makes predictions that put ARM in positive light.

Imagine my shock /s

londisteCISC vs RISC argument has been essentially dead since micro-ops became a widespread thing.
The success RISC-V has had in adoption has nothing to do with anything related to ISA or architecture. RISC-V is free. This is their winning argument :)

Can you explain the failure of this instruction: AVX-512? It turned out to be a power hog, and was abandoned.
Why was that? And what solution does RISC-V have to compute as an alternative?

john_... in case Intel's fab business fails and with it, Intel fails.

Not gonna happen.

john_Nvidia could use it's power in GPUs to push the ARM platform to gamers and enthusiasts.

NVIDIA has no interest in wasting effort on specialised products for "gamers and enthusiasts" when they can and are making money hand over fist from putting more and more powerful GPUs in datacentres.

john_There is huge support behind ARM this time and Apple already proved that moving from x86 to ARM, is not impossible.

It's easy when you have a handful of applications and don't need to support games.

well if AMD and Intel are still being complacent and (maybe) complicit while their competitor is gonna eat up their precious cake, then maybe they shouldn't be putting their legs on their corpo-desk and actually do something about it.

Windows had as much chance as going the way of the dinosaur by the 2030s. So it could be a moot point...

I don't believe Microsoft will get away with making each version worse than the last by then.

In fact it'll surprise me if the forced planned obsolescence requirements of Windows 11, doesn't get the ball rolling.

Only legacy x86 apps apps are keeping people on the ecosystem and they're slowing disappearing.

Dr. Dro, the one I'd buy right now is $INTC because it's in a low and it has the biggest growth potential behind NVIDIA (though $NVDA will be purchasable again after the 10-1 split occurs, can't spend $1280 in a single paper).

I actually did that last month... put few grands... Intel just cannot go bust... as a principle. X86 won't go away because it is good or whatever, but because of software and compatibility.

Take some most popular ARM platform, that is arguably raspberryPI... can you boot a PCIe video card? It it will take years, decades for something to budge there. PC definition is variety at core. You can have the hardware, but software will not catch up faster because something would look like a good idea.

Also don't put your eggs in one basket. I still cannot see why not to make low powered ARM devices. If the client wants we can do it, similar fashion Fuji did..

ARFCan you explain the failure of this instruction: AVX-512? It turned out to be a power hog, and was abandoned.
Why was that? And what solution does RISC-V have to compute as an alternative?

AVX-512 is not an instruction, it is an instruction set extension. This does not directly affect the core x86-64.

It is a power hog mainly because it deals with 512-bit operands. Longer registers, wider execution units. There are good use cases for AVX-512, the hardware implementation just was not there and the question became whether it is worth it. For these instructions to work the hardware implementation can be different. AMD started supporting AVX-512 in Zen4 by using two 256-bit floating pipes together and only in Zen5 by extending the length of floating point units.

What do you mean what solution does RISC-V have to compute as an alternative?

AVX-512 does not come from empty space. There was MMX, then there were versions of SSE and 3DNow!, then AVX, AVX2 and now AVX-512. Width of operations increased along with that as well. Aside from operations themselves also the amount of data being processed. IIRC SSE brought 128-bit, AVX2 256-bit (and heat issues in Haswell and Skylake), AVX-512 brought 512-bit. There are use cases - mostly related to multimedia - where SIMD and wider data processing is beneficial.

ARM isn't different either. There were SIMD, VFP (not SIMD but vector), Advanced SIMD/NEON extensions at one time or another. These are the successful ones that survived (or a version for them did). And these were not a standard functionality in ARM until some point (v8?). This type of operations are now being added to ARM proper (at a somewhat slower pace).

RISC-V will have to basically walk the same path once it starts to work with the same or similar data or algorithms. Wider operations, bigger units. At some point they cannot affort to keep on doing things by 64-bit chunks when competition does 2, 4 or 8 times wider (for example see what AMD did with AVX2 and AVX-512 support - initially made do with smaller units and used 2 of them to do the operation but then made the units bigger in the next iteration). Whether that will end up with more complex operations is an interesting topic and we will have to see. In theory they can keep the hardware simple but that will only move the same complexity to software.