What you wrote refers to ? Read again, i wasn't talking about gpus mate
As to what role will Nvidia play with ARM it is yet to be seen. I'm sure they can give Apple a hard time without breaking any law just for the fk of it, like delaying their schedule for 3 or 6 months due to ...made up transition reasons of their own.

You may not consider it but Win 8 and 10 mobile was bought by some people. Many tried a Lumia for daily usage although most of them quickly switched. Perhaps a 3% of the market (at some point) isn't impressive but it was a number they could approve upon, yet MS preferred to pull the plug on the whole thing.

Errr and how is this is news bro? They have been doing so for over 10 yrs now for their iphones+ipads. But as i said it's one thing to design a gpu to play 3 x 4k streams and another thing to design a gpu that can accelerate the creation of such content the way an Nvidia or AMD gpu (and soon Intel too) would. The leap from a light usage/multimedia playback purposed GPU to a serious professional one is gigantic. So where is this gpu please. I haven't seen one and chances are neither have you. They may succeed in making such a gpu but not anytime soon. So "Apple are already rolling their gpu" is not a thing until they can compete with the big boys, that is the Quadro and Radeon Pro workstation cards. A good example, Intel, has been rolling out similarly spec'd and usage oriented gpus for ages but up to now was never a competitor in the gaming or workstation market.

This not what i meant. Nvidia at any point can throw away the already established relation model with phone manufacturers ARM had, roll out a series of generic models, as they do with their gpus and indirectly force manufacturers to use one of Nvidia's designs instead of their own. It has happened with SLI x86 chipsets in the past and although they gave that up, history has a fame for repeating itself. Of course this cannot happen overnight, but seeing how all companies absorbed by Nvidia were diminished or just simply became all Nvidia tech, there maybe a chance to kiss away the friendliness of licensing an ARM cpu and be welcomed to Nvidia's world and the according pricetags.
And after all it's just a reflection of mine on things as they unfold man. i never said any of these thoughts are certain or facts but you reply like you have an issue with people thinking out of the box or at least that's how i take it. You may wanna participate, instead of stating the obvious we all know.
Cheers

 

of what I understand, Apple is planning to make own gpu solutions, so both nVidia and AMD are out of the game (within few years),
also upcoming Apple silicon Macs most likely won't provide PCIe or Thunderbolt, so there won't be a reason to bother with dedicated gpus either

 

I think there is a bit of confusion or crystal ball prognostication going on here. Apple's family of Ax Bionic chips are ARM derivative chips. There is more involved in a Central Processing Unit than the silicone printed diodes and resistors that make up the physical chip.

This is the point at which things get reductive and simplified, the current ARM Internal Set Architecture (ISA) and the availability of supporting processors and controllers required to build a motherboard with functioning peripherals at a price point makes transition to ARM from Intel feasible. Even if the silicone Apple produces isn't a clone of a generic ARM design and contains physical deviations, improvements and optimizations the language and logic the chip uses to communicate with the rest of the controllers, buses, peripherals and the way compilers and assemblers communicate with the CPU are constrained by ARM Reduced Instruction Set Computer (RISC) logic.

It's a much more daunting task to create an ISA and get partners producing supporting hardware to that logic than it is just making a chip. Think of it like USB or Thunderbolt, there is a standard and people make stuff to that standard so that when you plug it in, it works. Well x86, x86-64, SPARC, PowerPC, RISC-V are those standards. The economies of scale that are now dedicated to ARM RISC embedded peripherals and controllers make it possible to deliver a computer at or below the price point of an Intel based computer. There have always been competing ISAs and there is even the OpenSource RISC-V ISA from Berkley (the Same people that put out the OBSD subsystem that Apple machines and compilers run on) but there just isn't the availability of commercial hardware suppliers producing integrated controllers and peripherals necessary to support a price capped production of mass produced personal computers.

Mass produced personal computers are what we are all using, unless some of you guys out there are browsing AudioSex on exotic chipsets with bespoke supercomputers then you are the exception to the rule.

RISC-V might be an option in the future if it takes off and there is enough affordable bits out there to make it possible to transition to. For now though Apple is married to ARM and seems comfortable with the licensing model and what it costs. ARM could turn out to be just a stepping stone and Apple's displeasure with Intel over internal tensions, the loss of 5G supply, security concerns, thermal/computation/space efficiency and it's recent decline in prestige could have have made it a necessary stop gap to RISC-V, we'll probably never know. Maybe someone will write a book about it in 10 years. The only thing we know is that Apple is planning on transitioning to ARM64 ISA based cpus and that Nvidia, probably the only company that Apple probably hates more than Intel at the moment, has acquired and is now managing ARM. We also know that Nvidia, like Qualcom is a viscous and hyper-competitive company (not a bad or a good thing) that is willing to strong arm competitors and allies to the furthest limits of the law in order to achieve its goals. We don't know how that will play out in the future though, only time will tell.

This might mean that RISC-V Macs and iDevices are on the horizon faster than planned but the determining factor for that will be if other chip makers and device producers switch to and start supporting and producing IEEE-754 chipsets. The jump from ARM64 to IEEE-754 shouldn't be as big of leap as most co-processors already support it, but I digress as this is already getting too technical and boring for a music forum.

The short version is that we as consumers often fail to ask one simple question when we think about our ideal products: how much is it going to cost in terms of parts, prestige and human labor to achieve and are all the things needed to make it available. We as consumers though are always quick to complain about what things cost and how we are unwilling to pay for anything that isn't the cheapest and least expensive thing available right now.

We'll just have to wait and see.

 

Last edited: Sep 15, 2020

Refers to "I'd like to see Apple getting a license from Nvidia lol". Existing licenses are not affected until their expiry date. Nvidia don't even seem to care about mobile, they grabbed ARM to sell big boy chips to server rooms. Little iPhones don't interest them much.

Which is why Windows 10 on ARM is not "a thing" and can't "improve as well". They are too far behind at this point and they don't seem to be willing to change that. Microsoft's focus is not Windows, not Office, not Xbox, and definitely not ARM/mobile devices - it's Azure and cloud services.

They can, actually. A12X Bionic in 2018 iPad Pro already outperforms both Surface Pro 6 (with i5-8250U) a Dell XPS 13 (with i7-8550U) in Geekbench 4 (sauce). In video editing it transcoded a 12-minute 4K video faster than the 13-inch Macbook Pro (with i7-8559U):



All of this with over 40% longer battery life than the XPS 13.

Keep in mind this is mostly CPU & iGPU performance, as the laptops above are running Intel UHD 630 graphics from what I understand. That said, according to this sauce, it's just below a GTX 1060 (desktop) in terms of performance. It was shown playing Shadow of the Tomb Raider in 1080p @ at least 30fps in x86 emulation mode. The integrated GPU also outperforms Ryzen 4500U laptops in OpenCL benchmarks.

It's already a decent integrated GPU to the point of being comparable to a low-mid tier dedicated GPU. And they are yet to show a bigger version of the SoC dedicated to laptops, not tablets. So I'd expect a decent improvement there. They certainly aren't slaves to Nvidia or AMD GPUs. You're talking about Quadro and Radeon Pro, yet there are no Quadros shipped in Macs, and there's nothing Pro about Radeon Pro. "Radeon Pro 580X" shipping standard with the Mac Pro is not a Pro card. It's a rebranded RX580, which is a lower midrange card that's already getting dated. It's the same chip as used in 27" iMac - a OEM/mobile version of the RX580. With 8GB of VRAM. That's a $180 card brand new.

sauce

Can very well be, though at the moment Nvidia are already on the regulatory radar and anything monopolistic they try to do works against them. So I don't expect anything shady, at least in the short term.

Not really, just trying to balance things out with an opposite point of view. Sorry if that came across as being negative and crossing out anything you came up with, I just like to keep any replies direct and in a bullet point style to keep everything as on-topic as possible. Cheers!

 

In the meantime…

NVIDIA is about to cop ARM Holding for 40 billion

It's on now…

 

NVIDIA has said it will keep licenses and licensing intact. Yes, to Intel and Apple as well.

They know they have to because, well, anti-fucking-trust.

 

So, being slightly better than the integrated i-gpu in Intel/AMD gives any indication that Apple can build gpus for pro usage? With what, iMovie? I'd like to see an Apple GPU accelerating Final Cut for example, thanks.
Notebookcheck finds Apple's igpu slightly below the 1060GTX with 30fps @ SOTR on x86 emulation?? But a semi-decent 1060GTX will give over 80 fps @ SOTR in High Quality and about 50fps in Ultra Quality. So, by the numbers, this is either a bad comparison due to the overhead and resources taxing of the emulation itself, or a somewhat suspicious exaggeration. I mean wtf, my home's 1050ti with an i3 8100 gets about 60 fps in SOTR/High Quality (i have finished the game twice).
Btw, at some point they have to start giving their customers real solutions not indirect like TB3 connectivity or half cooked like a Radeon 580. And when i refered to Quadros and Radeon Pros i wrote "workstation cards", which of course the 580 on the MacPro isn't.

Neither do i, but is two years a long time or short term...

No offense taken mate, i enjoy the opposition. I understand now you try to keep it neat and i welcome your quality.
For what it's worth, i like to try and challenge myself with thoughts that have no basis other than speculation on possibilities. So you are doing no harm, although all the info is known, at least to me. Trying to foretell what will happen though, is a different case than the sum of all facts so far, that's why i am thinking out loud and you may excuse me for that, as i can tell it may come out as conspiracy theory or even childish sometimes. You be well
Cheers

 

Last edited: Sep 16, 2020

I like all you said, but why should we keep this thread out of the speculative please. You have any problems with prognostics? This is not the MacWorld mag. Not everything has to be proven fact. I mean even MacWorld and the likes, do speculate on possible future outcomes.
So, Nvidia is only second to Intel in terms of evilness hehehe. I won't sit around doing the usual flashback on this, people interested know well how they came to be. I will just post this once more, just in case some peeps missed it:

Cheers

 

@taskforce

I was just saying that in the near future Apple to ARM64 is going to happen because it's actually possible and that APPLE SoC are ARM ISAs whiteout any grey zone in which they can avoid an IP license.

I read some talk of them skipping ARM to head to exotic ISAs like RISC-V which would be nice but economically unfeasible since Apple is just not big enough to generate the necessary economy of scale to make things affordable.

By all means, pease do continue to to prognosticate, it's fun to guess the future. I just wanted to keep things somewhat rooted in the realm of possibility and the near future as I find this thread very informative and entertaining. I just felt things were already heading to 2050 too fast

As for inter-company vitriol in my mind in the soap opera/TMZ reality TV relashionship model it's always been...

Intel (ex-wife that was a post-op TG but kept saying in public she was on fertility treatments and that APPLE was shooting blanks because they wanted children but could not have any.)

Nvidia (gold digging ex-mistress that ran up all Apple's credit cards while she had an other nouveau rich boyfriend [Bit-Coin Farmer] who she was already planning to leave him for cause kash is king.)

Qualcomm (back stabbing BFF business partner who used your ideas and help to make their modems the best but then said they came up with it on their own and would not let you use your own ideas in your own products.)

Oracle (Just because why not)



If Steve Jobs was still alive and at the head of Apple the list would be different of course.

 

Last edited: Sep 16, 2020