Discussions with Apple regarding license agreement | Hacker News Reader

Discussions with Apple regarding license agreement

imgtec.com

254 pointsczr80192 comments

In the GPU space it is impossible to not infringe on the IP of other vendors.

In fact it is the major reason GPU vendors give for not having an open source driver. I have spoken to the CTO (Jem Davies) of ARM about the GPU drivers and open sourcing them more than once. And every time I've gotten the reply: "No, we can't, it opens us up to IP infringement suits."

Full disclosure: I used to work in the ARM GPU division.

One should keep in mind, this is true of basicly any software whatsoever. It ALL infringes on patents, nearly without exception.

There is always legal risk in open sourcing code. Infact there is legal risk from pretty much any action whatsoever. Good, responsible companies don't let that become a barrier to doing the right thing.

If ARM really cared, GPU stuff would be open source. The fact it isn't pretty strong indication they don't. Don't just accept it when lawyers say no.

After all AMD and Intel both have full open source graphics stacks, and the world hasnt exploded yet.

Should also be noted, keeping stuff closed source is not very strong protection against reverse engineering, if it was DRM would actually work, and it never really has, dispite decades of effort. So its doubtful that keeping stuff closed is much protection against patent lawsuits.

> If ARM really cared, GPU stuff would be open source. The fact it isn't pretty strong indication they don't. Don't just accept it when lawyers say no.

Easy for you to say, harder when millions to hundreds of millions are at stake.

Its always easier to be an armchair critic, but I bet there are more than hundreds of millions waiting for whoever can de-crappify the embedded ARM driver ecosystem.

Especially when this crap is at the root of why virtually every embedded linux product has serious security flaws that never get fixed. You cant update the kernel if you cant update the drivers.

If there is a reason on why ARM's GPU drivers are not updated on your device, it's hardly ARM's fault. ARM doesn't ship the SoC or the device directly to you, it ships it to the OEM. And it's the OEM who states the agreements.

If you want to blame someone, starting by blaming the OEM.

ARM still writes the drivers for the chip and then dosent update or open them up. Not much OEMs can do about that.

But ARM cannot, and won't, officially support the OEM's devices and start releasing updates for them. The process is a bit more complex than you might be imagining.

It still has to start at the top, and if they did update their drivers at least open source projects could then update on said platforms instead of being stuck on a 5 year old kernel till the end of time.

ARM has a reference design which includes drivers. The OEM is responsible for the final design and validation. ARM cannot take responsibility for validating 1000s of implementations that vary in both hardware and software configurations and validate them for billions of devices. Neither can the FOSS community. For the handful of boards which are popular enough to have sufficient tracking for the FOSS community to take the wheel if an open source driver was available the OEM has sufficient incentive to support their platform. For the other 4,999,999,990 LowHo NiHao industries unbranded boards it wouldn't matter in the first place.

> but I bet there are more than hundreds of millions waiting for whoever can de-crappify the embedded ARM driver ecosystem.

Hyperbole much? 7 billion people on the planet, and "more than hundreds of millions" of them are waiting with bated breath for the "de-crappification of the embedded ARM driver ecosystem"?!?

I think you'll find the hundreds of millions in question were dollars, not people.

You're almost certainly right. Oops!

Good, responsible companies don't let that become a barrier to doing the right thing.

Despite what Stallman would have you believe, open sourcing your own code is neither right nor wrong. It's just a choice.

Its most definitely a quantifiable, collective wrong when that choice leads to a total security disaster like the embedded ARM situation. It might not be so bad if they bothered to update their drivers, but they dont even let other people try to do so.

Open source software has plenty of high profile security disasters too. Pretending otherwise is either ignorant or disengenuous.

Unluckily this is true. But there is a central difference: If such a security bug occurs in an open source software, you can in principle look for the bug source yourself to fix it to secure your computer to against attacks. If it is closed source, this is hardly possible or often such a self-defense is even illegal.

In theory yes, but just to take one well known example, the HeartBleed SSL bug was introduced in 2012 but wasn't found until 2014.

At least it can be patched, if the equivilent bug occured in a closed source arm SoC driver, it would NEVER be patched.

Why wouldn't it be? If a vendor doesn't patch their drivers to fix security issues, people will stop using their products.

GPUs are hardware, not software. The patent situation there is much clearer. Critical hardware features of GPUs belong to different companies and the licenses for using those features often include highly restrictive licences for the code that drives them, or contractual obligations to use and not publish such code because doing so would reveal internal details of the hardware implementation of those features, which are trade secrets. Many drivers also load proprietary microcode into the CPU. AMD's 'Open Source' drivers do this.

Put simply, the GPU vendors do not actually own all the rights to their own driver software. AMD had a project to improve the state of their drivers on Linux, but the approach they took was to try and rewrite and open source the code implementing the public API to the closed binary blobs inside the drivers, to make it easier to maintain the stability of the drivers across kernel versions. Open sourcing the blobs wasn't possible for AMD because the code in them just doesn't belong to AMD.

>Don't just accept it when lawyers say no.

Which lawyers though? It's not necessarily ARM lawyers that are the problem. You'd need to get agreement from the dozens of patent holders of various bits of the technology, most of whom are unknown as their identities are confidential.

Open sourcing commercial products like hardware architecture and algorithms to the public is not exactly an easy thing to do.

There are millions of registered patents and the chances that your clean-slate ideas were already invented and patented are really high.

Open sourcing means exposing patent infringements to the public (even if you are not really aware that you are infringing anything), which means that you need to invest on a strong legal team in order to go through all possible patents and to deal with all possible litigations you might face.

In order words, open source requires much more than ideals, it also requires butt loads of money.

Why has open source been so successful on the CPU then? Branch prediction, say, is no less a patent minefield than GPU framebuffer tiling. Yet gcc and LLVM have no trouble shopping optimizers.

This is an excuse, basically. They just don't want to because they fear revenue lost to compatible implementations.

Those are compilers. What open source CPU is there?

Uh... the GPU drivers we're talking about in this subthread are precisely "compilers" for the shader architecture (and configuration generators for the texture units and framebuffer layouts, etc...). In fact for architectures other than NVIDIA's the hardware-facing part of the linux driver already is open source.

The only secrets left are the bits responsible for turning OpenGL (or Vulkan now, I guess) calls into programs to run on the GPU.

Wouldn't it be fair to say that GPU architectures (at least thus far) tend to change a helluva faster than CPU ones?

I can think of AMD adopting and then ditching VLIW for one.

OpenSPARC T1 and T2 are open source and don't have out-of-order execution.

That's completely different, there is interest to expose your CPU architecture in order to help people write better compilers and better programs.

But on a GPU, the whole interaction between the GPU and the application is abstracted by APIs like OpenGL and Vulkan and you own the driver, you own the compiler, you own the implementation and you own the architecture. So companies tend to protect their "secrets" since they own the whole product.

If you are asking me why is it different... Ask these patent trolls instead:

- https://techcrunch.com/2014/09/04/nvidia-sues-samsung-and-qu...

- http://www.anandtech.com/show/11101/amd-files-patent-complai...

It's not an excuse, it's the sad reality.

And there is no interest in exposing your GPU architecture in order to help people write better compilers and programs?

Just as on a GPU, the whole interaction between the CPU and the application is abstracted by APIs[1] like C and Python.

[1] Yeah, we don't call them APIs. But then we don't call GLSL a "API" either. They're all languages.

I have heard that ARM has not been too pleased with IMG buying MIPS and competing directly with ARM.

Perhaps ARM would license their GPU patents to Apple and provide some IP protection.

Apple could also license patents from AMD. They don't compete directly.

AMD sold their mobile division to Qualcomm; Adreno is just Radeon with the letters rearranged. They may not be able to get involved in mobile.

Can we ever fix this system and get a better balance of rewarding innovation vs. accepting patent abuse?

How about this for a litmus test. Present a problem to new 4 year college graduates that they haven't heard the answer before. If x% of come back with a solution, that solution is obvious enough to be consider invalidating the patent.

Or let validity be judged by independent groups of practioners who weigh the novelty and also the impediment of a patent to world progress.

As those who've written patents already know, they are not even decent scientific or technical documentation. They're mostly a bunch of jargon and phrases meant to tic off legal checkboxes. No one would ever write a document that way if the sole purpose was to explain something.

First off, that problem already exists in the hardware, the damage is already done (as i stated in my blog entry http://libv.livejournal.com/26635.html under "the patent excuse").

There was another big smoking gun recently, at linaros conference in budapest. Here finally some truth was spoken at 9:40 into the video: https://www.youtube.com/watch?v=dPjSm_98rgo&t=580s

Like IMG, ARM is not selling just the hardware design, it is also selling driver development services. When the driver is open source, and competitive, everyone can sell such services, and there's a lot less of those services to sell. ARM has revenue depending on this.

This is one good thing about the IMG revenue coming from Apple, it was mostly licensing costs, not services. If IMG had been selling services on top, their revenue would've been reduced by 75%+ instead of "just" ~50%.

> In the GPU space it is impossible to not infringe on the IP of other vendors.

It's high time to reform the corrupted and broken patent system. Yes, I know, like having non-corrupted politicians, this is not going to happen.

Yes, that is why pretty much All GPU markers have deal cross licensed with each other.

The story from last year was that Apple's price for IMG acquisition weren't high enough, and Apple asked if they could buy ONLY the graphics portion. Which resulted in a NO from IMG.

Compared to ARM, some estimate Apple is paying $10M for the ISA and roughly $0.1 for each Ax SoC. So that is ~30M per year, ( There will be other cost for processor other then Ax, such as the Motion Processor etc ) Compared to paying roughly $80M USD for the Graphics alone, which Apple doesn't even use their design, nor their services ( Software Drivers ) they provide. So what likely happen was Apple wanted to pay far less, ( $20M? ), but their New CEO wanted them to pay for more.

My guess is that in the end Img will relent and license whatever Apple wanted for a price. If not, I am not sure how Apple will deal with old games that wont work with their new GPU which does support PVRTC.

This is worded like this because ImgTech had a duty to disclose the notice they received from Apple.

Check out what the share price did: https://www.google.com.au/search?q=LSE%3A+IMG (down 60% - looks even more dramatic in the 5-day view)

Does the decrease in share price make them a more attractive acquisition target for Apple?

I wouldn't put this past them as an acquisition tactic, Imagination drops MIPS, Ensigma and suddenly they look very attractive just for the patents alone.

Yeah, this is an open letter informing the world that imgtec is turning from a company producing useful stuff into yet another patent troll.

Patents belong in the garbage.

What happens when you put all your eggs in one basket..

Wouldn't that be considered manipulation? Or at least, using inside information.

I don't think it would be insider trading as they would not be buying or selling stocks based on non-public information.

OK, I guess. But it would have been illegal if Apple had shorted IMG before informing them about the end of licensing, right?

It makes them a more attractive acquisition target for anyone who wants a decent IP portfolio.

I think apple already has shares in IMG. This would be huge blow to them since they are their 1 big customer.

Bingo.

"Apple currently holds more than 8 percent of the company."

http://www.theverge.com/2017/4/3/15158982/apple-iphone-gpu-i...

A few years ago I was working for a company which provided value added mobile services to the UK market. The biggest money maker at the time was premium SMS, however we didn't have our own SMS gateway - we just resold services of another provider (which was unbeknown to clients at the time).

Shortly after I joined the bosses decided that we were giving away too much money to the provider, so began talks for an acquisition. The provider was willing, but were asking for much more than we were offering, so the talks fell apart. As such we ended up building our own SMS gateway and switched all our traffic to that - exactly replicating their API, bugs and all, in less than a year.

As we were their main customer they ended up losing 70%+ of their revenue overnight when we switched. A couple of years later we bought them, for much less than we originally offered. Both companies at the time were private companies, but I wouldn't be at all surprised if the same happens here.

As soon as I started reading this comment I was thinking "I worked at a company that did the exact same thing!" Ha ha, small world :)

Almost certainly not.

Apple just announced to the whole world that they are no longer interested in Imagnation Technologies' IP.

Announcing to the world that they were interested in Imagination Technologies after all (by acquiring them) would not only have made the prior announcement deceptive, but also (almost certainly) be considered market manipulation, given that the announcement caused a 60% drop in the stock's price.

You're likely correct overall - However worth adding in this case it was Imagination making (and phrasing) this announcement, not Apple.

Imagination had little choice if they received notice from Apple, as noted by an other commenter

> UK regulations on stock markets are quite tough, publicly listed companies must share information that could impact the company as soon as they can.

Yup. I'm making a different point.

There is a long history of customers using their leverage over a supplier to say that you either allow us to buy you, or we take our business to a competitor - or in house.

There is definitely a lot of money to be made here, can't see it staying so low.

Why not? They don't seem to have any other customers.

Indeed. From their 2016 Annual Report, "There are no parties with whom the Group has contractual or other arrangements which are essential to the business of the Group except the contract with Apple Inc."

https://imagination-technologies-cloudfront-assets.s3.amazon...

Page 30 for those interested. I was surprised that it was actually in there, phrased that way.

Great time to buy!

The last time their stock was this low was eight years ago...!

This could make the company a pretty attractive acquisition target for a competitor?

Haaaaang on.

When Apple started designing their unibody Macbooks, they needed to put a hole in the aluminium for the power LED. There was only one company in the world with the equipment to make those (multiple-of-nanometers-wide) holes.

Simple solution: Apple just bought the whole company. Voila, now Apple can cost-effectively produce those holes with one less renewable contract.

This one might get interesting...

---

EDIT: With the above being said, note this nearby comment: https://news.ycombinator.com/item?id=14022436

> EDIT: With the above being said, note this nearby comment: https://news.ycombinator.com/item?id=14022436

Also there were rumors about buyout talks last year, Apple commented that they "had some discussions with Imagination, but [did] not plan to make an offer for the company".

Also if they get to repatriate their overseas money that would allow them more ability to acquire companies important to their business.

I love how passive aggressive this is – it shifts from "it's a shame Apple is leaving" to "we could sue them over it" pretty quickly towards the end. They imply it's difficult to do, but also that it's basically already complete... bizarre post to make public.

The undertone certainly seems like they'll be paying a lot of attention to Apple's compliance of their patent portfolio. Given how much of their revenue comes from Apple I wonder if they'll have to become an enforcer of their current IP rather than generator of new IP.

This tone is what you get when your CEO hasn't slept in 3 days because of anxiety.

And what's with the "This announcement contains inside information"? at the end? Once they release a public statement, isn't everything in that statement then <not inside information>?

So that everyone who is handling this news before it is released (i.e. the publisher) knows that it contains sensitive information and may not be acted upon.

Time for ImgTech to go Linux/Mesa friendly maybe? With an open source graphic stack they could become a great choice for projects like Raspberry Pi and other hw manifacturers: imagine a Vulkan-ready linux laptop that combines a rpi Compute module with the graphic potential from their GPUs... Hope they would be this brave.

I think they're too attached to their "precious IP" to allow that.

It seems their closed source drivers are full of bugs and incomplete functionality

According to the post at https://www.phoronix.com/scan.php?page=news_item&px=Power-VR... from 2015, PowerVR is/was working on an open source driver. Does anybody know whether something materialized from this?

Last time I pinged their spokesman (Alex Voica), the answer was that they hadn't found someone with the 'right fit'.

When was that?

> Apple license fees and royalties, as disclosed in Imagination’s Annual Report, represented revenue of £60.7 million for the year ended 30 April 2016 and are expected to be approximately £65 million for the year ending 30 April 2017

> Imagination believes that it would be extremely challenging to design a brand new GPU architecture from basics without infringing its intellectual property rights.

Well Apple has a ~$75m per year incentive to do just that, so I Imagination better hope they've really got a stranglehold on all the necessary IP.

And they better not infringe on any Apple IP as well (or IP that was acquired)...

Apple can build a GPU without infringing ImgTec patents, through a patent agreement with another GPU maker(including ARM, so they're pretty much covered).

Perhaps not. Existing GPU makers protect themselves through cross-licensing their patents: you can use mine if I can use yours.

However, it sounds like ImgTec is considering going full-on troll, so they won't need access to the other patents any more and can just sit under their bridge extorting tolls.

Patent trolls are non-practicing entities. Imagination Technologies is hardly a non-practicing entity.

Sure, today. But they're "concentrating on the licensing business".

> Imagination believes that it would be extremely challenging to design a brand new GPU architecture from basics without infringing its intellectual property rights

Can't believe they would try to use this as an argument in their own favour. It's either (yet another) proof that the patent system is broken, or that their patent are as invalid as can be.

> It's either (yet another) proof that the patent system is broken, or that their patent are as invalid as can be.

Why did you assume the patents would be invalid? It is possible that they hold patents fundamental to the field so every other GPU make has to (cross-)license the tech. There is also no obligation for FRAND licensing on GPUs, to my knowledge.

That was a rather oddly worded press release. Much more defensive than I would have expected and quite arrogant towards the end. They make it sound as if no other company on earth could do what they do which as we know does not usually work out well for those who claim such a thing.

It may be more understandable when you know that two dozen of their employees have left for Apple since 2015 (some highly placed), they're basically saying they believe Apple's internal GPU effort piggyback on Imagination's tech.

So Apple's not even bothering with buying the company for the talent? Apple is just straight up poach talent, and cutting out the company?

The people don't belong to the company, why should the company get "a cut"?

The people are free to go, but getting the same people to develop the same product again without infringing patents is a challenge.

Conversely, if there are any group of people who will be able to develop the same product again without infringing patents, it's the people who wrote the patents in the first place.

Patent violation doesn't have a mens rea component - either you violate the patent by doing exactly what is described in it, or you don't. If you change even one aspect, you're no longer violating the patent. I'm sure the people who wrote the original patented algorithms can easily think of ways to accomplish the same thing that don't violate it.

This is very true. However, they could still lose a court case if they failed to prove that the design they shipped benefited by information that was confidential - even when the information was about what "not to do". Still, this is probably among the hardest case to bring on and win for a company, imho.

Patents aren't confidential information. In fact, that's kind of the whole point of patents.

Well, not really so odd if your stock drops 77% to a 8-years low in a single morning.

Judging by the tone used in that press release, they are in deep trouble. It does not even matter that £60.7 million is a small portion of their business. It could be the opposite of what happened to ARM when Apple decided to put their chips into Newton.

Their 2016 annual report notes £120.0m revenue (90 from IP/royalties) and £136.5m in operating expenses with an operating loss of £87.0m. Their new CEO mentions refocusing the company on IP and royalty income. Now Apple, which pays 60% of that income, is ducking out. Ouch.

In the same report, they mention "There are no parties with whom the Group has contractual or other arrangements which are essential to the business of the Group except the contract with Apple Inc." All your eggs in one basket...

What's funny is in their financial risk assessment they focus on bullshit like investment shifts and outside IP threats, but gloss over the whole "what if that huge contract our company depends on happens to fall apart?"

The discussion between Apple and these guys must've really gone south for a nasty and desperate press release like this

> The discussion between Apple and these guys must've really gone south for a nasty and desperate press release like this

That's what I found most intriguing. Why go public with this? (Especially with ongoing litigation)

Signalling a potential lawsuit so that Apple enters into a patent licensing agreement.

UK regulations on stock markets are quite tough, publicly listed companies must share information that could impact the company as soon as they can. Public opinion is pretty susceptible to this too due to events in recent years.

I call this good old fashioned and fair to shareholders. Not tough. Wall Street is basically where you get the worst , and unfair treatment as a shareholder.

Hinting at possible future litigation (and ultimately winning damages from Apple) is them trying to offer some light at the end of a very dark tunnel.

>> What's funny is in their financial risk assessment they focus on bullshit like investment shifts and outside IP threats, but gloss over the whole "what if that huge contract our company depends on happens to fall apart?"

Perhaps thats why they've refocused on IP, projects that are actually profitable.

Apple accounts for one half of their revenue.

http://www.reuters.com/article/us-imagntn-tchnlgs-apple-idUS...

In my opinion, IT reaction is not wise. No matter if this comes after acquisition negotiation, or not, who is going to engage with IT in another long term partnership with such level of reaction? (not sure if it could be qualified as bullying or not)

This really isn't news, it's been known for years. A GPU team is based in Austin (knowing Apple, they likely have multiple).

They've also had a load of job adverts out for hardware graphics positions in London over the last few months. I think I read their new team is supposed to based there (but I can't find that source now).

This doesn't come very surprising to me, as Apple has been in talks to acquire Imagination Technologies but has decided not to: https://arstechnica.com/apple/2016/03/apple-acquire-imaginat...

I don't know the reason, but I suppose it is the same reason why they decided to build their own gpu independently. Apple also already own(s/ed?) a part of Imagination Technologies.

The rumors about Apple working on GPU cores internally long predate the Imagination "talks" (which were apparently rather shallow). In fact significant Imagination employees started leaving for Apple since at least 2015: Dave Roberts (senior design manager) left in October 2015[0] and Jonathan Redshaw (VP hardware engineering) left in November 2015[1]

[0] https://www.linkedin.com/in/dave-roberts-9ba859a6

[1] https://www.linkedin.com/in/jonathan-redshaw-30400990

Exactly. Maybe they figured that they already hired the key personnel, so there was no need to buy the whole company any longer.

If Apple takes advantage of the depressed stock price to turn around and buy them, it would not be the first time.

https://en.m.wikipedia.org/wiki/P.A._Semi

Haha, and then Apple would own MIPS.

iOS has always exposed patented imgtec texture formats to app developers. For backwards compatibility Apple will have to license the formats. At least until the relevant patents expire.

This is death throes of imgtec.. and as a side effect, put a fork in MIPS, it's done.

> Further, Imagination believes that it would be extremely challenging to design a brand new GPU architecture from basics without infringing its intellectual property rights, accordingly Imagination does not accept Apple’s assertions.

I know that designing new GPU from scratch isn't an easy task - but is it THAT hard that they don't trust that one of the richest companies in the world is capable of doing it?

Intellectual property is complicated - there must be some patent that Apple would infringe anyway. Perhaps the one on breathing...

Would is be possible to patent some "biological respiration and oxygen replenishment" process and start suing babies?

You might be able to make it stick in East Texas.

Or PowerVR texture compression, which according to Anandtech is the canonical method for iOS development:

"Because Apple’s SoCs have always used GPUs from the same vendor, certain vendor-specific features like PowerVR Texture Compress (PVRTC) are widely used in iOS app development"

[0] http://www.anandtech.com/show/11243/apple-developing-custom-...

They could easily force change on the app for such change. Look at what they're doing regarding Bitcode and how quickly it might become the default.

It's orders of magnitude harder to statically check for at app submission time than Bitcode vs. Native is.

The keyword is without infringing its intellectual property rights. Sure, they certainly can design a GPU, but design it in a way that does not step on the toes of others? Especially when you learned from them? That's the difficult part.

My understanding of the matter is the Apple GPU has had much of it developed in house for quite a while anyway (both on the SW and HW side). The only reason they didn't let go of Imagination earlier was presumably because they relied too much on their IP and patents.

Now I guess Apple have judged they no longer need Imagination and patents or more likely they are able to navigate the legal consequences successfully and the likely fines are less than the savings made (or worth it for the control Apple gains).

This announcement has been a long time coming and isn't particularly surprising.

There's also the possibility that Apple knows it infringes on certain patents but have done a risk assessment saying the lawsuits would be cheaper to go than to buy the company outright.

Going with the lawsuits also mean a good chance some patents could get invalidated but if not, Apple could end up paying more in the long run. Given how much money they have, they may not care because they could just buy the company outright later.

And if it is that necessary to use that IP to create this technology, even on a basic, generic, level, is the IP really valid?

This is interesting:

"Imagination believes that it would be extremely challenging to design a brand new GPU architecture from basics without infringing its intellectual property rights, accordingly Imagination does not accept Apple’s assertions."

Could Apple not be using someone else's IP, for example from NVidia or AMD?

If licensed to do so, that's how ARM works as far as I've known.

AMD does not license to third parties, neither does Intel. Options (ignoring acquisition and design from scratch) are:

* NVidia's Tegra graphics cores, possibly (not clear whether they block-license let alone architecture-license to third parties)

* ARM's Mali, definitely licensable

* Broadcom's VideoCore, possibly

* Vivante, definitely licensable

* Qualcomm's Adreno, formerly Imageon, unlikely

Also note that at least two dozen Imagination employees left for Apple between 2015 and 2016 http://appleinsider.com/articles/16/10/13/apple-poaching-gpu...

I was under the impression all the big modern consoles were using customized AMD GPUs. For example, according to the PS4's Wikipedia entry [0], its GPU is a "Semi-custom AMD GCN Radeon (integrated into APU)". Are those customizations made by AMD for Sony, instead of licensing the technology?

I found an article claiming Intel is licensing AMD GPU technology [1]. If that's the case, would it be surprising to see Apple do the same?

[0] https://en.wikipedia.org/wiki/PlayStation_4

[1] https://seekingalpha.com/article/4042977-new-details-intels-...

Yes, those customizations are made by AMD. They aren't even very custom, fail0verflow have managed to run a lightly modified linux kernel with the open-source AMD drivers on PS4 hardware with full graphics acceleration[0].

[0] https://fail0verflow.com/blog/2016/console-hacking-2016-post...

As far as I know this are more or less custom configurations of the same CPU i.e. how many CPU/GPU cores, how much cache etc.

> Are those customizations made by AMD for Sony

Yes, same for the Xbox One, they both use APU customised by AMD.

I have a theory that when you're the size of Apple, new licensing opportunities will open up.

>AMD does not license to third parties, neither does Intel.

If the money was right, they would. Apple is not a random third party. Heck, it could buy AMD with a year's spare change.

AMD did, however, sell Adreno to Qualcomm when money was tight after the ATI purchase.

And looking back it wasn't a smart decision. Adreno=Radeon

That may be true but unlike Adreno, you know Apple does not resell any of the patents to anyone else, so it always stay contained to Apple's products. Given how much revenue they can take, it's a very tempting business agreement.

Along with the other things they sold that don't seem so smart now. Like abandoning the low power x86 Geode processor in 2009. Interestingly, according to this article [1] they sold Adreno for only $65 million, less than IT is making from Apple each year in royalties.

https://www.xda-developers.com/a-look-back-at-amds-history-a...

Hold up. I thought AMD was actively courting third parties to partner with to do mixed IP development (wrt graphics and GP GPU compute) and that licensing restrictions that they had were almost entirely related to the x86 side of their business.

I remember rumblings about that some years ago (circa 2010~2012), but I don't remember anything ever coming of it.

Only 3 of those provide high-end ip.

  * NVidia Tegra - possible
  * ARM Mali - possible, but would not allow for much differentiation
  * Qualcomm - has never sold its GPU IP seperately

Nvidia's Tegra is key to their mobile compute efforts. There is no way they are going to license it to anybody. They would sell Apple the chip... maybe. But they would want Apple to buy into their whole stack, which Apple isn't going to want either.

As near as I can tell, the CPU in Tegra chips are just vanilla core designs that Nvidia licenses from ARM. The secret sauce is the GPU. Given enough money I think Nvidia would license their GPU designs to Apple as it would expand the reach of CUDA and make it more of an industry standard.

As far as the CPU cores go, yes, there isn't much to distinguish them there.

There is no way in the foreseeable future that Nvidia will license their GPU designs to anybody.

The only way you'll see Nvidia IP on an Apple box is if Nvidia-made chips are in there, but Nvidia will require buy in on the rest of their software stack. But Apple won't want to cede that much control of their platform to Nvidia.

Nvidia already had tremendous reach, they don't need Apple.

AMD does license to third parties, at least for x86:

https://www.extremetech.com/computing/227059-amd-announces-n...

From further down, "The company’s...IP...includes the key processing blocks needed to create the SoCs...that power all mobile, consumer and embedded electronics." I've never heard of the company but my assumption is that they've patented some critical part of the assembly process which means Apple is relying on them.

nVidia made similar claims against Samsung (and Qualcomm) a couple of years ago. A lot of folks speculated that Samsung would lose not only because nVidia is a leader in GPU business, but also because Samsung is a foreign company.

Long story short, nVidia lost. (https://www.theregister.co.uk/2016/01/07/nvidia_and_samsung_...)

Samsung initially asked Qualcomm to deal with it since nVidia was really going after Qualcomm's Adreno GPU used in Samsung's smartphones in the US. Qualcomm instead decided to sit back and twiddle their thumbs as nVidia filed lawsuits against Samsung (and Qualcomm). Many predicted Samsung would end up like Apple vs Samsung considering Qualcomm's political clout and jury bias in the US, but the USITC quickly put an end to Qualcomm's misadventure. Samsung countersued and won, and Obama couldn't be bothered to reverse USITC's opinion, impending import ban, against Qualcomm.

Yeah, if you think about it companies like Apple have very little reason to pull out from a licensing deal unless they have a very good reason to (aka non-infringing tech on the way). They're not going to pull out and come back and say "We're sorry. Can you give us the same rate you offered before?"

No, I can say with certainty that they are working on their own mobile GPU.

They have been hiring a lot of graphics people and putting a team together.

And another thing that most people are not really aware: Apple had a lot of saying in the architectural and design decisions of Imagination's GPUs that ended up on their iPhones. A good part of the development actually happened at Apple's offices with Imagination people flying over.

So they know what they are doing, they are very well familiar with Imagination's GPU and they are more than capable of developing their own thing from scratch.

> Apple had a lot of saying in the architectural and design decisions of Imagination's GPUs that ended up on their iPhones. A good part of the development actually happened at Apple's offices with Imagination people flying over.

So that's why Imagination is insisting Apple can't not infringe: they know Apple won't have a cleanroom implementation not using the guys who've talked to Imagination.

Apple have a classic "they saw the copyrighted sourcecode" problem on their hands.

Not really, you seriously think Apple would just let ImgTech guys come in without lawyers and agreements and all of that sort? Apple has extensive experience in this area, they had ImgTech signed everything possible to protect Apple and to indemnify themselves. It is a risk that ImgTech also took by allowing Apple deeper into the development process. This isn't a one-way street here.

Apple is extremely potent in protecting its technologies. There is no way they just let random ImgTech fly in and work on stuff with them without any agreements in advance. If this happened, ImgTech is going to be an easy billionaire by the end of the lawsuits they could do.

While I have no doubt that Apple works closely with their hardware partners by flying their engineers in to work on projects, I seriously doubt it was as simple as the OP made it sound.

It actually is as simple as the OP said.

That was the main reason why Apple went for Imagination instead of ARM or Qualcomm when it comes to mobile GPUs.

Imagination market cap has been falling hard in the latest years, so hard to the point that their only customer until now was Apple. They were desperate and they signed very risky deals in order to keep Apple as a customer.

And Apple is a complete control freak when it comes to their products, the idea of not being able to control the stuff they put on their products is unthinkable to them.

So Imagination signed a bunch of architectural deals (instead of purely implementation deals) because that was the real product Apple was looking for.

Don't let yourself be mistaken, this whole situation is far from a surprise to Imagination. They knew this day would come, they were just trying to cling on to the little market they could find until they found another deal in order to stay afloat.

I'm not sure where we crossed wires but we're saying the same thing, nothing you said changed what I said.

What I meant by OP is that it is not as simple as flying their partners in and they start working together and then leave. Apple doesn't just do that without ensuring everything that happens stays in Apple only. So, flying ImgTech guys in and out does not mean ImgTech owns the patents to what they did at Apple want or the other way around, Apple can ensure they have the exclusive rights to it.

Patent threats and trolling should be punished, one way or the other. "We are the authority for GPUs, if you want to design one from scratch - it's impossible" should be punished.

> Apple has not presented any evidence to substantiate its assertion that it will no longer require Imagination’s technology, without violating Imagination’s patents, intellectual property and confidential information.

Sounds like IMG has some fundamemtal IP that nVidia and AMD already license? They certainly have tied up their Tile-Based Deferred Rendering method, which never sounded that efficient to me (yet, Apple beats others graphically, so I probably don't appreciate it - maybe to do with cache efficiency?).

Apple, having understood this technique very well, might have thought of something even better.

They certainly would have to be pretty sure about their tech before sending this notification. Let's hope this... betrayal doesn't have the same fallout for customers as ~~google~~ maps did.

I already made a comment about Apple's GPU here: https://news.ycombinator.com/item?id=14021814

Apple actually has its own mobile GPU, built from scratch.

The Tile-Based Deferred Rendering GPU is an advantage from the GPU sharing the same memory with the GPU. On a normal desktop GPU you need to transfer huge amounts of data from the main CPU RAM to the Video Ram but on a mobile device, the CPU and the GPU sit both on the same memory system. This allows you to architecture your GPU in a different way.

Both ARM's Mali and Qualcomm's Adreno use Tile-Based Deferred Rendering.

Both Nvidia and AMD have already switched or are moving towards tile based rendered to reduce bandwidth (and hence energy) costs.

Source? That's a huge architectural change for them.

http://www.realworldtech.com/tile-based-rasterization-nvidia...

So that's tile based rasterization not tile based rendering. There's a huge difference. This is basically just a cache before the ROPs write into DRAM.

EDIT: Also, I don't think his test is proving what he thinks it is.

Looks like Apple was nearly half their business/revenue. The company's market cap (after this morning's crash) is only $275M, which is less than 5 years worth of royalties...

Interesting that Apple was that much of their revenue. Reminds me a little of PortalPlayer from the iPod days, though it was much more dependent on Apple (90% of its revenue). But I remember the rumor mill echoed Apple was unhappy with them. Eventually, Apple decided to switch MP3 chip suppliers and the PortalPlayer imploded.

http://www.eetimes.com/document.asp?doc_id=1160843

Keep in mind this is not just hardware but software related.

Right now I have IT graphics tools to optimize 3D models for mobile devices (for example).

For us developers, it means optimizing to a different standard, while worrying about backward 'optimization compatibility'.

Plus we'll need new software tools for development on the new GPU.

I imagine the new GPU will support the same Metal API. As this is probably something that Apple has been working on for a while now I'd assume that Metal was designed with this new GPU architecture in mind. How much of those tools would need to be updated to a new GPU assuming the current API was specifically written for it ?

But even if the API is the same, it doesn't mean it will have the same performance? Something that was not a problem with previous GPUs could become a bottleneck and vice-versa? And since you need to support both GPUs or else you drop older iPhone support, it might get annoying.

I don't really know if that would be the case though, I don't know anything about GPUs.

>But even if the API is the same, it doesn't mean it will have the same performance? Something that was not a problem with previous GPUs could become a bottleneck and vice-versa? And since you need to support both GPUs or else you drop older iPhone support, it might get annoying.

If "in certain cases, the performance profile can be annoyingly different" is the outcome of a tectonic shift in the underlying hardware, it would be a significant and praise-worthy achievement.

But it is Apple, one of rare "experienced" companies that can pull this off well (not perfect but enough). Apple has done two (or three?) complete CPU architecture changes and developers just had to recompile their app. Remember fat binaries from the PPC>Intel switch over?

I suspect as long as the graphics tools use Metal, it will be a quick transition.

Looking the trend of doing everything in-house, it is not a surprising or unexpected move from Apple's part.

Apple certainly has the resources to devote themselves to doing this right, if they really want to...

Silly question, why can't Apple just buy the company outright.

It is not a silly question, it's just that no one outside of Apple and/or ImgTech knows.

The most logical reason would be that Apple has all it needs when it started building its own custom GPU many years ago and ImgTech has no patents it needs, which is entirely possible despite what ImgTech has said.

It is also entirely possible that Apple knows it has a few patents it is infringing and because on its last several years of experience, it would be cheaper to go through a lawsuit than to license. The reason? By going with the lawsuit, you're likely to get some of the patents invalided and royalties drop. It is possible it could go the other way, Apple paying more but it is extremely difficult to provide willful infringement when Apple can provide evidence of its own custom work that started before ImgTech was used in a clean room environment. But that's also easy to mitigate if ImgTech shows Apple hired the key personnel that worked on the same technology as the personnel should've known it was patented.

Around 2011, IMG turned down acquisition offers from Apple and Intel.

I wonder how likely it is for Apple to support Vulkan on its hardware.

Considering they do not allow Vulkan on OSX when there is hardware that supports it about zero.

Possible but given the way mobile hardware works I wouldn't hold your breath.

They won't want devs to miss out on hardware features because the Vulkan abstraction (by its nature, cross platform) is missing a feature.

So it may happen but I'd imagine only after a strong native toolkit exists already

They have already decided to go for their own API, Metal.

You know, for this news to come out along side the news from Imagination Technologies Group, you know they have to have something impressive already.

So when is apple going to merge cpu and gpu? Less than 5 years or more.

Merge in what way? They are already on the same die.

For low power mode I would expect them to remain separate.

Conceptually.

But CPUs and GPUs are "conceptually" different: they are designed from the ground up for opposite workload types.

Does this mean cheaper iPhones?! (huehuehue)

Ummm No. It probably means more money on Apple's cash pile in Ireland.

And possibly the end of one of the only open /and/ synthesized CPU.

No way I have implied that... -.-

Finally a software company makes a GPU.

Everybody who has programmed in VHDL or Verilog knows that building hardware is very close to writing software actually. It's just a level of abstraction lower, and there is a lot more parallelism, but the concepts are very similar.

Hahahaha, No.

Synthesizing ASICs and FPGA designs from Verilog/VHDL/insertnewhighlevellanguagehere may be close to software, but I guarantee you that creating high end production chips with such strict efficiency constraints guarantees that you need to do heavy simulations, and most importantly, has massive verification (it is 10/7nm after all....) efforts.

Ok, let me rephrase it as: it is like writing software with real-time constraints.

Writing HDL is quite similar to software, but the "stack" you deal with is entirely different. It's like saying that writing embedded C is similar to front-end JS (hardware interrupts vs. events, external world vs. DOM, input voltage vs. user input, etc.).

It's more like the difference between writing C and writing a cookbook. As in a kitchen cookbook.

Especially if your goal is efficient, synthesizable HDL!

I have described hardware in VHDL, and I've written a fair bit of code, and those processes are not similar at all.

Apple is a software company now?

"This announcement contains inside information."

That single line at the end of the document means that any buying or selling of Apple or Imagination shares could be considered insider trading.

Is this legal black magic to void trades that have hammered their stock price?

I blame Microsoft for Busines' current infatuation with Imaginary Property. Human civilization was built on the sharing of ideas and the ability to run with your neighbor's smart plan. Now you get sued for patent infringement and the neighbor expects to take rent for the rest of his natural life, and that of his undead estate, for what is almost always, at some level the synthesis of prior human endeavor and ingenuity.

It's fucking bullshit. Luckily the Chinese gets this and wipes their collective arses of what is essentially a western fiction.

I agree with your overall sentiment, but Microsoft didn't start the patent trouble, or even the extreme overreach of copyright law; they were the victim, not the aggressor, in the first round of such.

Don't blame Microsoft – they're no stranger to the game but this has been exploited since the Supreme Court created software patents as a category. Here's a reaction to flagrant trolling back in 1993:

https://www.fourmilab.ch/autofile/www/chapter2_105.html

Prior to that, you had companies like IBM, who were famous for using their hardware patent library as a club to get competitors to sign licensing agreements.

Microsoft hadn't really even started when Apple itself had already perfected the fine, dirty art of submarine patents. We're using about early 90's, if not late 80's.

Reply on news.ycombinator.com