Apple ends GPU contract with Imagination (PowerVR)


vcoleiro1

Hardcore Member
Joined
Jan 23, 2011
Messages
4,685
Apple has used GPUs designed by Imagination (PowerVR) in their SOCs for a long time now - ie those used in iPhones and iPads. . But it seems Apple are now going to design their own GPU for upcoming SOCs to go into future iPhone and iPads. This has led to a 63% drop in Imaginations stock price in one day.

Apparently Imagination is looking to take this to court.

Imagination also designed the PowerVR GPU in the Omap5 .

Over the years the SOCs that use Imagination GPUs have dwindled to almost being exclusively for Apple these days. So this obviously would be a huge hit for them, thus the 63% stock price drop in a day


https://www.engadget.com/2017/04/03/apple-is-building-its-own-gpu-for-the-iphone-and-ipad/



..
 
Last edited:
Well maybe they stop only giving a crap about Apple now. From what I gathered all of their support effort went there.
 
So Apple pretty much destroyed a whole company out of nowhere because they could save a few pennies on iPhones they already sell with a ridiculous profit margin?

65908913.jpg
 
The only device I have that is running a PowerVR GPU is my Dreamcast (I guess the Pyra will join that soon however) :p If Apple did this out of the blue it is indeed a "dick move" they should have maybe warned imagination some time before pulling out.
 
Apparently Apple owns 10% of Imagination and had tried to buy it out a year ago. After that didn’t go through, Apple allegedly poached some Imagination engineers.

Now you have this happening. I wonder if this will end with Apple buying Imagination for a discount. The more cynical might say that was the plan .
 
Ditnt run much games on my Iphone 6S so far, maybe a bit Patience, and Snake Classic,
Much time i use my Iphone as an "Ipod" and as Internet Browsing and Telephone Device (and to make some pictures)..
For Gaming, i rather use Switch, New 3DS and Vita, and Pandora/Pyra,

But maybe Apple whill somethime make an Iphone whit Gaming Controlls like the Xperia Play..
 
Sounds like even Apple is sick of insufficient GPU Documentary and poor driver support. ;) Maybe PowerVR will learn it's lesson now.
 
Sounds like even Apple is sick of insufficient GPU Documentary and poor driver support. ;) Maybe PowerVR will learn it's lesson now.

I know I'm "too close" to be considered a remotely impartial observer, but I *really* don't understand these comments. Are you moaning about the lack of public documentation about the PVR architectures? If so I can assure you our customers get them. And again, vested interest, but my own personal development stuff have hit a *lot* more problems on other vendor's driver stacks than PVRs. Even nvidia, who people seem to think is the "Gold Standard" of opengl.

At the end of the day, our customers are *not* individual devs. There's no real difference for pretty much any other vendor in the same mobile space (unfortunately, though I *hope* if one vendor starts releasing more public docs/code and making it a "selling point" everyone else will jump on board). It just feels like PVR get a *lot* more abuse for the "same" things everyone else in that market is doing in exactly the same way.

It gets extremely frustrating being constantly insulted for something we can't see in ourselves. Perhaps we're completely blind. I'm sure you know better.

Plus, apple used their own driver stack. Always have.
 
At the end of the day, our customers are *not* individual devs. There's no real difference for pretty much any other vendor in the same mobile space (unfortunately, though I *hope* if one vendor starts releasing more public docs/code and making it a "selling point" everyone else will jump on board). It just feels like PVR get a *lot* more abuse for the "same" things everyone else in that market is doing in exactly the same way.

Please do not take personal offense from random comments. We appreciate any and all support that PVR can officially provide and any and all support that can happen unofficially without offending PVR. Most of us are thankful for having an SoC at all.

Yes, we all wish it were more open. It is a long road from the traditional closed source model with it's safety of, "everyone else does it this way," to, "open source". There are likely contracts and licensing with other 3rd party companies that would have to be taken into account before even considering moves in that direction.

Projects like this are in a bit of a doughnut hole though. PVR provides the documentation and driver support to TI. To get the hardware at a discount, I understand that the Pyra project has little to no official support from TI. (I understand that it does get some off-duty assistance though - thank you all involved.) Developers on the Pyra aren't all even employees of DragonBox. In theory, every user of this device will be a potential developer on/for the device. That decentralized model makes the lines of accountability and liability very fuzzy and all the more difficult for PVR via TI via DragonBox to supply documentation or code references through to the underlying systems.

So, thank you for any and all assistance you are willing and able to provide. Most of us understand that there are limits. Please don't blame us for wishing for the 'more' that simply can't happen without miracles from the executive suite.
 
Too bad Apple doesn't make (yet) toilet paper, I'd enjoy wiping my *ss with one of their product.
 
Serves Imagination right. That's what you get for solely relying on one greedy corporation and refusing to document their GPUs.
Now that Apple became traitors, hopefully this should make them reconsider their support for Linux/Android.

From what I understand they DO document their GPUs and provide this to the entities that they perceive as being their customers. In our case that would be to TI.

Maybe, if we ask nicely, they might be more accepting of the concept of providing more support for Linux/Android end users, but that could be impossible or at least impractical. After all, you're asking them to release what could amount to trade secrets or patents that they are in turn licensing from other parties. The whole financial and legal infrastructure of these things can be far more complex than you are likely taking into account.

More information and more support for the things we care about (Linux drivers) isn't going to be made any easier if they perceive Linux users as being insulting and unappreciative.
 
Please do not take personal offense from random comments. We appreciate any and all support that PVR can officially provide and any and all support that can happen unofficially without offending PVR. Most of us are thankful for having an SoC at all.

Yeah, I'm not trying to say "people shouldn't be encouraging public docs/source release", but the "Bad, IMG suxxx!" comments are exactly the opposite way of doing it, being not remotely constructive and only possibly harmful (as in I find it insulting, and you're less likely to get someone who you've insulted to help you get what you want). It feels like it's aimed at the people who are in the best position to help.

I (personally) *want* to work on open source stuff myself. I think it's a massive shame we can't release stuff. I personally think OSS is a massive benefit, and it has significant value to me (IE I pretty much only buy hardware that is supported by OSS drivers myself, and am willing to pay *more* for that). It's a massive shame you should feel "thankful" for second-hand support and SoC offcuts. But money talks.

What *should* be done is to drive demand at the customer level. If TI (or whoever - TI is probably not the best example as they've closed down a lot of the SoC division and currently running of 'old' gpu designs so no new license fees :) started seeing demand for this, and they believed they could sell more/charge more per unit/some other benefit of such things and feed that back into the GPU vendors sales team that would be a a *much* more useful. There's a large amount of FUD and 'possible risk' to releasing docs and code. Beancounters currently see no benefit to counteract that possibility. So the default answer will always be "no" until that is changed.

The current sales model in the mobile/embedded GPU space simply doesn't fit with products that aren't being developed by a single vendor for the consumer market. Arguably intentionally, as there is (currently) little visible profit to such things.
 
What *should* be done is to drive demand at the customer level. If TI (or whoever - TI is probably not the best example as they've closed down a lot of the SoC division and currently running of 'old' gpu designs so no new license fees :) started seeing demand for this, and they believed they could sell more/charge more per unit/some other benefit of such things and feed that back into the GPU vendors sales team that would be a a *much* more useful. There's a large amount of FUD and 'possible risk' to releasing docs and code. Beancounters currently see no benefit to counteract that possibility. So the default answer will always be "no" until that is changed.

The current sales model in the mobile/embedded GPU space simply doesn't fit with products that aren't being developed by a single vendor for the consumer market. Arguably intentionally, as there is (currently) little visible profit to such things.

There is a lot of 'chicken vs egg' in that though. It is difficult to drive sales demand for open products at a price premium when the components needed to make them are not available with open options.

Where this could open up is if a company similar to TI could come up with a product line from inception to production with the declaration of being entirely OSS. The contract with the GPU provider would then be negotiated with those expectations including expected driver support and the sales price of the resulting component would reflect those upstream costs & risks. Unfortunately, without an existing and established customer base for that kind of product, they would have to make a leap of faith - which isn't something that large corporations are likely to do if they can avoid it.

The success of the RPi has shown that there is a significant base of Linux on ARM customers - but at a very low price point. Would those potential customers be willing to purchase higher end products at a premium price point? The presence of the Intel Cherry Trail has 'capped' the price range for this type of endeavor. It's hard to sell a $50 ARM SoC with the market presence of a $40 X86 option. So, to 'get there', the ARM SoC would need to be a superior end consumer product. Mere clock cycles and cores on the CPU are great for embedded systems. On the consumer product end, though, the GPU is where it would truly need to shine. And that brings us back to the chicken and the egg...
 
Back
Top