tmay

avon b7 said:

tmay said:

avon b7 said:

tmay said:

avon b7 said:

tmay said:

avon b7 said:

tmay said:

elijahg said:

cloudguy said:

dk49 said:

If ARM has its own AI engine now, what does it mean for Apple's Neural engine? Is it possible for Apple to completely discard ARMs AI engine in their processors or they will have to build theirs on top of ARMs? If yes then will it not break ARM's licence? 

See above. Apple is an ARM Holdings co-founder. They have a perpetual architectural license.

The fact that they were an ARM Holdings co-founder is no longer relevant to anything, they sold (AFAIK) all their interests in ARM long ago, and have never had a controlling stake. The perpetual architectural license was acquired when Apple bought PA Semi in 2008. NVIDIA also has an architectural license, along with MS and Qualcomm, so there's no reason they couldn't match Apple's M1 CPU speeds - except their engineers aren't good enough, apparently. So not really sure what Nvidia sees in ARM.

A disadvantage of not controlling the whole stack, as Apple does, is that Qualcomm, Samsung, et al, haven't control of Android OS, or Windows OS, and so will never have SOC's as optimized for end users, as Apple will for its own ecosystem. It may not make all that much difference in a mature marketplace, though it appears that Apple still sees an increasing user base, and still benefits from its tight integration of all of the technologies that is has at its disposal.

Android vendors can control the 'whole stack' if they wish to. Android is open source. It would take a huge investment and vendors would have to effectively re-invent many wheels but it's an option. There advantages and disadvantages to both scenarios. 

Huawei has been forced to do exactly that, and as a result, perhaps it could be argued that it controls as much, or more, of the 'whole stack' as Apple.

It's 5G modem and WiFi chipsets are designed in house, for example. Apple's aren't.

It can also 'optimise' the stack beyond the CE boundaries of Apple, as it also produces Cloud hardware and services along with AI hardware and services. It also develops it own battery and charging technologies. Not to mention participating in and designing the core communications technologies that are the backbone of today's modern day devices. Apple devices included of course. 

In software, GMS is being replaced by HMS etc. 

Qualcomm and especially Samsung are also well placed to do the same if necessary but for different reasons, they have no need to.

It's worth noting that Huawei has also been reportedly greenlighted to use ARMv9 and just like all vendors, has the option of using RISC-V too. 

You would be accurate in stating that Huawei is on a possible path to that, but factually, Huawei is not anywhere close to where Apple is in SOC development, and that takes into account the short term advantage that Huawei has with integration of 5G into its smartphone SOC. I also take exception to their Android fork being very developed, though of course, you have disagreed with that in the past, against published evidence that it is mostly a vanilla copy.

Samsung has attempted to, but has never been able to replicate Apple's success and continue to trail in SOC performance, and density, and Qualcomm develops a range of SOC's for its many customers.

As I have stated before; every year, Apple ships about 70% of its units based on its single, current (A14), A series processor, and this year, it appears that Apple will approach 250m iPhone units, which is in the neighborhood of 175 million A14's, not including iPad's. Not in anyone's universe will Qualcomm come close to those numbers for the Snapdragon 888, nor Samsung for the Exynos 2100, nor Huawei for the Kirin 9000. 

What's interesting is how much of an advantage Apple continues to carry over its competitors, and the M series is yet another instance.

Huawei is right up there with Apple on SoC development. On process node and even besting Apple on transistor density. On timing. On getting their modems on SoC, on ISP and DSP development, on secure enclave etc.

The amount of processors Apple ships is totally irrevelant. What does that have to do with the advantages of owning the whole stack? And you are making some utterly wild projections anyway. Why not try to run with something that is more realistic, like what they shipped last year? Either way, quantity would still mean nothing but if that is what you want, Mediatek reportedly shipped over 350 million processors last year.

As for the advantage Apple continues to have over competitors, are you claiming that not being able to ship a latest generation 5G modem on SoC is an advantage?

And as for 'the Android fork' (eh?) not being very developed, it already contained more APIs than Google GMS Android, and two days ago, Beta 3 was released, adding 1,000 more APIs. Considering no one outside China has even seen the system, you are jumping to A LOT of uninformed conclusions. Especially as the system already has a release date for handsets: 24 April.

I think we should just wait and see on that one. 

Huawei/HiSilicon is dead in the water without access to a leading node fab, so what you state is factually untrue. Rumors of a Kirin 9010 at 3 nm are pure fantasy, as there are no fabs in the PRC that are capable of less 14 nm. There were very few Kirin 9000 produced, perhaps 8 million total, and the Kirin 9000 was trailing A14 and Qualcomm 888 in performance.

Apple owning its own stack is why Apple has the M1, and will follow up with more performant M Series. Apple owning its own stack is why Apple doesn't have to be concerned with SOC cost, and hasn't created mid, and low range, variants of the current A Series, other than AX which powers iPad.

I misstated that Apple doesn't have 5G integrated on the SOC, That should be restated as Apple does not have 5G integrated on the A14 die. Of course, in the future, Apple will integrate 5G on die, but in the meantime, consumers aren't concerned about how Apple implements 4G, and units sold in proof of that.

Yeah, we should "wait", but it should be noted that you have no specific knowledge of Huawei's Harmony OS as it pertains to smartphones either; you're merely parroting Huawei Marketing.

Anyone (Apple included) has a problem if they have a product design that cannot be manufactured. 

Not long ago, the entire industry was affected by a very similar problem but that wasn't politics, it was COVID-19.

A Kirin SoC on 3nm is anything but 'pure fantasy' . Logic should tell you that if it weren't for politics TSMC would actually be testing  that processor with Huawei right now. Each Kirin chip is basically on a five year design window.

But none of that has even the slightest to do with what I was replying to: the advantages of own the 'whole stack'.

Ah yes, Apple and the M1. You did slip that in there but then Huawei has MindSpore, DaVinci, Ascend (which goes from earbuds up to AI clusters with 1024 cores), solutions for Airport and Port control, robotics, Power solutions and a very long list of etc. 

The whole stack you say? Why not. It makes sense for some companies but not necessarily for others. 

This is 'old' now but I still haven't seen an Apple equivalent. Have you? 

https://e.huawei.com/en/products/cloud-computing-dc/atlas/atlas-900-ai?utm_medium=psm&utm_source=corp_facebook&utm_campaign=HC2019&source=psm_corp_facebook

As for HarmonyOS, I am definitely not 'parroting' Huawei Marketing.

The information I provide is publicly available and comes from the head of HarmonyOS software development (not marketing) . Your claims, and this is ironic btw, do seem like the product of pure fantasy.

If I say it's best to wait and see what is released, it is because there is nothing official yet. That's reasonable on my part but you seem to be able to reach conclusions that are based on... what exactly? Definitely not anything Huawei have actually said.

Oh right, the ARS article! Is that it though? Nothing else? 

Did you bother to contextualise that article? To see how those conclusions fit in with what Huawei is actually claiming? Did you question anything? 

Where are those 16,000 APIs hiding for example? Being a multikernel system with a kernel abstraction layer, how many kernels were looked at?

How is it possible that one of the world's largest home appliance brands (not handsets) has already confirmed support for Harmony OS on a whole suite of products?

https://www.gizchina.com/2020/11/11/midea-launches-its-first-batch-of-products-running-on-harmony-os/

Of course HarmonyOS is already out there either in version 1 form (TVs, cars and smartscreens), or parts of it working 'undercover' in watches and routers etc and doing things that Android simply cannot do. 

Xiaomi is able to purchase the full line of Qualcomm, so that avenue may be available to Huawei, but I doubt it. 

"A Kirin SoC on 3nm is anything but 'pure fantasy", but logic tells me that it will never exist, so perhaps it is you that needs to stop fantasizing, and since Huawei isn't going to get even a semi-custom processor anytime soon, then Huawei absolutely lacks the "full stack" for its smartphones. I don't think that Samsung will be able to save them either.

Thanks for playing.

Looking forward to Google's White Chapel SOC, so they can take advantage of the "full stack" as well, because of course there are other options besides Apple in the world.

https://jamestown.org/program/semiconductor-scandal-a-concerning-backdrop-to-xis-pursuit-of-core-technologies/

https://cset.georgetown.edu/wp-content/uploads/Khan-Flynn—Maintaining-Chinas-Dependence-on-Democracies.pdf

You have managed to tie yourself up in knots again.

Yes, Xiaomi has has access to the full range of Qualcomm products but what does that have to do with anything?

Huawei does not have the same access. You don't have to 'doubt' that because it is crystal clear and the result of a last minute change by Trump.

Samsung is a different story but there are only rumours at this point.

You have skirted mostly everything I challenged you with by either running up some new alley or outright ignoring it but for good measure you throw China into the soup.

I'm not surprised.

And 'logic' cannot tell you a 3nm Kirin will never exist when, at this junction in time, the only restriction in place is 100% political. There's a field where you should never say 'never'.

As for 'owning the full stack', perhaps you see things more clearly now than from your Apple centred earlier posts. Yes, the world includes more than Apple and clearly there are stacks, and there are stacks, LOL.

And in the spirit of Jaws...

"You're gonna need a bigger stack!". ;-) 

avon b7 said:

tmay said:

elijahg said:

cloudguy said:

dk49 said:

If ARM has its own AI engine now, what does it mean for Apple's Neural engine? Is it possible for Apple to completely discard ARMs AI engine in their processors or they will have to build theirs on top of ARMs? If yes then will it not break ARM's licence? 

See above. Apple is an ARM Holdings co-founder. They have a perpetual architectural license.

The fact that they were an ARM Holdings co-founder is no longer relevant to anything, they sold (AFAIK) all their interests in ARM long ago, and have never had a controlling stake. The perpetual architectural license was acquired when Apple bought PA Semi in 2008. NVIDIA also has an architectural license, along with MS and Qualcomm, so there's no reason they couldn't match Apple's M1 CPU speeds - except their engineers aren't good enough, apparently. So not really sure what Nvidia sees in ARM.

A disadvantage of not controlling the whole stack, as Apple does, is that Qualcomm, Samsung, et al, haven't control of Android OS, or Windows OS, and so will never have SOC's as optimized for end users, as Apple will for its own ecosystem. It may not make all that much difference in a mature marketplace, though it appears that Apple still sees an increasing user base, and still benefits from its tight integration of all of the technologies that is has at its disposal.

Android vendors can control the 'whole stack' if they wish to. Android is open source. It would take a huge investment and vendors would have to effectively re-invent many wheels but it's an option. There advantages and disadvantages to both scenarios. 

Huawei has been forced to do exactly that, and as a result, perhaps it could be argued that it controls as much, or more, of the 'whole stack' as Apple.

It's 5G modem and WiFi chipsets are designed in house, for example. Apple's aren't.

It can also 'optimise' the stack beyond the CE boundaries of Apple, as it also produces Cloud hardware and services along with AI hardware and services. It also develops it own battery and charging technologies. Not to mention participating in and designing the core communications technologies that are the backbone of today's modern day devices. Apple devices included of course. 

In software, GMS is being replaced by HMS etc. 

Qualcomm and especially Samsung are also well placed to do the same if necessary but for different reasons, they have no need to.

It's worth noting that Huawei has also been reportedly greenlighted to use ARMv9 and just like all vendors, has the option of using RISC-V too. 

"This was always something Broadcom did well. I recall many conversations with their executives who were proud of the fact their chip design libraries were portable, and they could make them at whatever foundry they saw fit. Qualcomm is similarly executing a dual-source foundry strategy as they have versions of the same chipsets made at both TSMC and Samsung. 

Companies that dual-source will be extremely well-positioned to weather a number of different storms that could come their way. From the geopolitical that I have outlined before, national economic issues, global catastrophes, etc. While this isn’t discussed as much publicly for obvious reasons, it is top of mind for many executives in the supply chain and those whose companies make products via semiconductor foundries."

I remember when Apple used TSMC and Samsung for production of the same A9 SOC. The response from the tech community was to compare and contrast the versions. I can see why Apple will never do that again.

https://arstechnica.com/gadgets/2015/10/samsung-vs-tsmc-comparing-the-battery-life-of-two-apple-a9s/

"And that's just what some iPhone 6S and 6S Plus buyers have run into. Using an app that has since been pulled from the App Store, some users were able to determine which chip individual iPhones were using and found that the phones with Samsung chips had significantly lower battery life than the phones with TSMC chips in certain tests. The findings got enough attention that Apple offered a rare comment on the situation, claiming that the test being used wasn't representative of actual use and that in "real-world usage" the difference between iPhone models with any combination of components was no more than 2 to 3 percent."

dewme said:

tyler82 said:

mac_dog said:

GeorgeBMac said:

Maybe they could send them some batteries -- or ideas on how to build an electric grid.

I’m sure they have plenty of ideas, but they don’t mean shit if it stalls at the government level. Both sides are so focused on politicizing everything, they forgot who they’re actually supposed to be working for. 

Texas is Republican-run. In the 1930s the state disconnected from the national power grid so that energy companies didn't have to be regulated. These companies saved money by not winterizing their natural gas, coal, oil refineries, and windmills. These things operate just fine in northern countries like Canada, Norway, Finland, etc. But the Republican-run "don't mess with Texas" independent boot-straps utopia is now begging the federal govt for handouts, the same federal govt they continually turn their back against. Now their arrogance and stupidity has come back to frost-bite them.

Putting the politics aside, what you said about the winterizing is right on. I’d also bet that many homes in Texas use heat pumps for heat, which revert to pure resistance heating coils (basically a giant toaster) in extreme cold conditions. This would put a huge load on electrical demand.

The Texas city of El Paso actually did heed the warnings of the failure to winterize their equipment in the decade since Texas suffered a similar series of extreme (for TX) cold related power failures ten years ago and only experienced a tiny number of power losses over a brief time period this time around. More short term thinking at play here. It’s sad that the people of Texas had to suffer due to the ineptness of their leaders, but unfortunately that’s the way things go.

Apple’s charity, along with similar efforts by other organizations and individuals, to help people in-need is certainly to be commended.

jdb8167 said:

commentzilla said:

jdb8167 said:

MacQuadra840av said:

jdb8167 said:

These still won't be processors for the Mac Pro or iMac Pro. Those are probably coming next year.

The Mac Pro will be using Intel chips for a few more years because Apple's M processor can't beat what is in the Mac Pro as far as memory, storage, features, and performance, especially the ability to run multiple VM environments.

Second, the iMac Pro is dead.  Apple hasn't done anything with it for four years.  It will likely be discontinued when Apple releases an iMac with hopefully a much better processor than the low end M1 chip.

Apple has repeatedly stated that the Apple Silicon transition will be over in two years. Since the latest start of the transition would be on the release of the M1 Macs, then logically Mac Pros are going to be announced by November 2022. I expect that we will see a Mac Pro replacement in early 2022 and it will beat the current Mac Pro in performance and match it for RAM, storage and other features. I also expect it will support discrete GPUs.

If the Mac Pro is the only model that supports 3rd party GPUs in any form that's going to be a really tiny market for graphics card makers. Is it even worth supporting? So it begs the question, could Apple simply scale up their GPU, Neural Engine, etc, units to replace them? Granted it probably wouldn't be on the CPU die but rather on a 5nm GPU co-processor tapped into the shared memory pool. Who says the unified memory concept cannot be extended to the very top of the line?

What's clear at this point is that Apple is cutting INTEL and AMD out fo their MacBook and iMac development, so why not the MacPro? Why would AMD bother to develop for a single Mac?

There are rumors that Apple is developing a discrete GPU. Is that easier or harder than developing drivers for high end AMD GPUs? Seems like developing drivers in cooperation with AMD might be easier and it solves a big problem for the really high end pro systems.

I'd imagine that Apple would have to negotiate with AMD to get them to support an Apple Silicon Mac Pro and/or iMac Pro. That might entail funding the driver development or some other incentive. Who knows? Again, there are rumors that Apple is going to create their own GPU but there is no way to know. One thing you can be pretty sure of though is that they developed the latest Mac Pro knowing that they were transitioning to Apple Silicon. So they have something in mind and I doubt it is to end of life the Mac Pro after a single hardware generation.