Why didn't Intel choose Nvidia for their integrated GPU? Why did they have to go with AMD?

Why didn't Intel choose Nvidia for their integrated GPU? Why did they have to go with AMD?

Other urls found in this thread:

tomshardware.com/news/amd-intel-x86-cpu,7285.html
twitter.com/NSFWRedditGif

It's for Apple.

Because Nvidia and Intel hate eachother

nope

$$$$$

>Why didn't Intel choose Nvidia for their integrated GPU?
No one is retarded enough to actually work with NVidia.
>Why did they have to go with AMD?
The only GPU company to operate semi-custom unit.

AMD has always been good with getting others to integrate there graphics since there more cost effective, Intel is still stupidly far behind in graphics and they want to compete with nvidia

nintenderp did.

>nintenderp did.
Only and only because AMD offers no tablet SoCs worth mentioning.

Because they didn't get a custom unit, it's a straight up Tegra no changes or anything.

That's not custom.
That's off the shelf abandonware

And most importantly: Why can't Intel into GPUs?

Because patents and AMD and NVIDIA have 20 years of experience in driver hacks.

Patents, softwarem engineers.
Qualcomm had to buy Imageon and pouch it's team from AMD to get a start in GPUs.
Intel has no such luxury.

PowerVR was up for grabs, they could have gotten some very efficient designs.

>PowerVR was up for grabs
Meh.
>they could have gotten some very efficient designs.
They've used their designs for Atoms back then.
It was meh.
It's either ATi or nVidia.
ATi now belong to AMD, and leather company will never sell itself to Intel.

>leather company

i don't get it. explain

nVidia.

Everyone hates nVidia, everyone.

Nvidia doesn’t produce SOC GPUs, while AMD does and has for a while.

This, and the primary reason is nVidia's astounding greed.
Intel used to license GPU patents from nVidia, but when the deal was up for renewal they asked for so much it was cheaper to license from AMD, even after considering Intel competes with them in CPU space too.

Because it would be a breach of their patent licence with Imagination Technologies.

Intel hasn't been using Imagenation anything for ages.

This.

I think he might've meant "latter". Either that, or he'd enjoy the company of someone in leather.

t. leather jacket man

What about ARM, Vivante, imgtec, and think silicon?

Shit, shit, shit and shit?
Especially ARM.

>these mobile gpus can't compare to my desktop gaymen rig
>they must be shit

They ARE shit.

and they sold their arm gpus to qualjew if Im not wrong?

Yes, Imageon along with it's team were sold to Qualcomm.

nVidia had market on machine learning,deep learning,autonomous car,HPC and supercomputer.

Intel, buy some ML/DL company,buy for a lot one autonomous car tech company,make processor and Compute units to HPC,Supercomputers.

Nvidia becomes big competitor of Intel, Intel want damage income sources like Laptop Gaming, to make nvidia grow slow.

>nVidia had market on machine learning,deep learning,autonomous car,HPC and supercomputer.
They are borderline irrelevant in HPC and supercomputers, like only 30% of Top500 uses GPUs.
Meme learning and self crashing cards will crash.
So, Intel is doing the right thing and is hitting the right market.

nVidia grow massive on markets, new supercomputer USA use IBM Power 9 and nVidia Volta.

>nVidia grow massive on markets
Literally meme learning bubble that will pop in a year or two.
>new supercomputer USA use IBM Power 9 and nVidia Volta.
There's only two, Summit and Sierra, and the first exaflop machine will be built by Intel, not nVidia.
Besides, silly jap thingies are sitting on top of Green500 and not nVidia's GPUs.

Because they said they don't care about laptop gaming

I'm actually curious what would have happened if AMD and Nvidia merged.

That was on the cards at one point, but Hector Ruiz and Jen Hsen Huang both wanted to be CEO.

So the deal fell through and AMD bought ATI instead.

intel hates nvidia.
nvidia faggots will NEVER EVER get an x86 license.
EVER

If AMD and Nvidia had merged, Intel might have bought ATI.

AMD would've died in 2010.
All hail big dies on barely yielding node!

What if Nvidia buys Via?

they got scammed pretty well by novidea

Probably because AMD has better integrated graphics than Nvidia? Nvidia is King of dedicated graphics but their integrated graphics suck shit

Intel is not using AMDs iGPU.

that's not how licenses work

...

Yes it is.

I have a feeling that AMD are doing this as a strategic risk, and are aiming at Nvidia.

I'm a little concerned about Ryzen mobile - but these are 45w parts, while all the announced RM chips are aiming for 15-25w, though maybe AMD will announce their own high end ryzen mobile chip soon.

Speaking of which, Broadcom might buy Qualcom - would that render the no-compete agreement they signed with qualcom null and void?

It's merely a way to put more Radeon into the laptops much the same way they put more Radeon into consoles.

They sold Imageon.
Sold.
It's not licensed.
The bad thing is that Broadcom might kill Centriq and ARM servers will go back to the stone age again.

I know, i'm saying as part of the sale they signed an agreement with qualcom to not compete in the mobile ARM market.

If qualcom ceases to exist as a company upon being purchased by Broadcom, would that agreement become void, or will the agreement pass to the companies new owners?

AMD needs a foothold in the 15-25 mobile market (AKA ultrabooks and normal laptops), and this is a great way to ensure that they have a presence in the high end while keeping nvidia out and the mid-low end (ryzen mobile).

I have no idea.
Besides, why would AMD bother with mobile GPUs?

Regardless of whether they buy an Intel or AMD CPU, they'll buy an AMD GPU.

Because, in the words of Linus Torvalds: "Nvidia, fuck you!"

The Tegra X1 is an excellent chip for a proprietary portable gaming system, but only for a proprietary gaming system and nothing else.
It's not running x86, so it can't be used for tablets, and it's not optimized for Open GL ES2.0, so it's not suitable for android or IOS.
Thus why Nvidia tried that shield thing so fucking hard, yet probably didn't managed to even pay for the R&D costs of the chip.

Nintendo most likely sent a team of Rick Harrisons to nvidia to get a retarded cheap price for the SoC.

Well duh, it's either Nintendo is getting them for peanuts, or NVidia will have to write off the invetory and kill the design.
NVidia had no choice.

Doesn't matter even if Intel gives them the x86 license. That is worthless.

AMD will never ever ever ever give them the x64 license.

And then there's a ton of extensions.
And then there's NVidia's morbid incompetence when it comes to CPU cores.
Its literally the only company that managed to develop housefire Transmeta CPUs.

nvidia shield used android though. its a mediocre arm chip with a strong igpu. Its kinda pointless in the smartphone market. Its only reason to exist was gaming tablets, which were never a thing. Thats why nintento used them

i dont really think that imagination technologies have better gpus than intel. They could buy them for their designs and probably manage to make something noteworthy in 4 years from now.

via still makes x86 64 cpus.

Apple's A# GPU (Imagination) >>>> Shitdragon/Tegra X1 GPU >>>> Atom GPU
In mobile Imagination was far ahead. Intel's laptop tier GPUs might be better at the expense of TDP.

Apple's current GPUs are not exactly Imagination.

you dumb mate?

Intel can terminate VIA's license if Nvidia bought them. They were about to terminate AMD's license because of GloFo.
tomshardware.com/news/amd-intel-x86-cpu,7285.html

Adding on, x86 licenses are non-transferable in the first place.

they jewed out on getting all the dirt cheap tegras nvidia couldnt sell so they could make more money

Apple's GPUs are not Imagination. They were made by Raja before he left to join AMD again.

They were Imagination before A10, they still are in some way.

It is because of cross x86-licensing and Vega/Navi architectures are better suited for APU applications.

This is what AMD RTG's long game is all about. Just wait until APUs start becoming "good enough" that they makes majority of discrete GPU solutions pointless. Just exactly what happened to discrete audio card solutions when integrated audio solutions started becoming good enough.

Vega does everything well assuming it works properly, and it doesn't right now.

it doesn't what?

Vega doesn't work properly right now.

The problem you're ignoring is heat. Yes, maybe when gate size shrinks enough and there are insane reductions in power utilization you will see this but not ATM. Take a look at the heatsink on GPUs and how huge it is. Take a look at how much power they draw, no think of putting that right next to a house fire CPU. Maybe for mobile this will work but desktop is not changing anytime soon. They have problems cooling CPUs and GPUs separately much less together

A10 and A10X are still imagination. A11 is probably at least 50% Imagination but we won't know until Chipworks tears it down.