From ATI to AMD back to ATI? A Journey in Futility @ [H]

Even if Zen utterly stomps Intel in every segment theyd still be lucky to increase market share into the upper 20% range. Half a loaf is better than no loaf

Intel iGPUs have never been premium parts and lack on the "perf" portion of per/watt. At the very least these will be more power efficient than Intel CPU + discrete GPU (in example the Macs with the RX460)
Well that would be the most bizarre business strategy of AMD, spend a fortune of their debt developing Zen/APUs and then just give up before launch because they will not gain 20% market share :)

I think the bigger question is who and which division is driving this deal in AMD with Intel if it is happening.
Cheers
 
I dont think everyone is looking at this the right way. If AMD makes a deal to be the gpu for every Intel cpu then they will have 100% of the integrated market. Last time I checked that is what most marketing people dream of and will make AMD quite a bit of money. Now they would like to sell you their apu over a Intel one, but either way they are making money on it. That kind of money will give their R&D a nice boost. If AMD sells a few less Zen apu chips they will be fine with it, they just wont make quite as much on each sale as they would on the Zen apu.
 
Kyle is getting some credit for moving the market a bit, I doubt this rumor is why AMD is surging but some sites are reporting that.

Mr. Bennett may have some info on a deal, but this has nothing to do with PC. This is an AI/machine leaning/ Autonomous driving project. Intel is attacking that market, they need a GPU to integrate into certain systems and products.

FYI RTG is an internal AMD organization meant to better manage R&D, teams and tech and IP. It has nothing to do with being spun off or being a different company.
 
It seems clear to me (I may be way off) that this GPU deal may be real, but would largely be intended for the Machine Learning, AI, Autonomous Car markets. Sure it may trickle into PC one day but that is a big undertaking considering that AMD GPU's are built to integrate with AMD CPU's (construct/cat/zen) You can't just plop into a Intel CPU without incredible amounts of work.

No way this can take place on current socket, a refresh of architecture coming at 10nm for Intel would need to take place before this could ever work on the PC side.

That said Intel could get behind ROCm and or begin to integrate Radeon compute solutions into the mentioned markets to combat Nvidia's push into data centers and the machine learning environment with their GPU and ARM cores.

Kyle is getting some credit for moving the market a bit, I doubt this rumor is why AMD is surging but some sites are reporting that.

Mr. Bennett may have some info on a deal, but this has nothing to do with PC. This is an AI/machine leaning/ Autonomous driving project. Intel is attacking that market, they need a GPU to integrate into certain systems and products.

FYI RTG is an internal AMD organization meant to better manage R&D, teams and tech and IP. It has nothing to do with being spun off or being a different company.
Maybe, maybe not. :)
 
I dont think everyone is looking at this the right way. If AMD makes a deal to be the gpu for every Intel cpu then they will have 100% of the integrated market.

From what Kyle said previously, Intel only needs patent protection from AMD, not iGPUs.
Intel current patent deal with Nvidia didn't put a single Nvidia iGPU inside an Intel CPU.
 
From what Kyle said previously, Intel only needs patent protection from AMD, not iGPUs.
Intel current patent deal with Nvidia didn't put a single Nvidia iGPU inside an Intel CPU.

The rumor has shifted a bit and now talks about a MCM design. Based on something I have heard as well I think this will be the very likely outcome. Intel has been laying off engineers in the gpu department, dont think they would be doing that if they just wanted patent protection. This lets Intel focus solely on the cpu and just place orders for the gpu which will save them money since R&D is so expensive. But it's still just a rumor until its announced.
 
The rumor has shifted a bit and now talks about a MCM design. Based on something I have heard as well I think this will be the very likely outcome. Intel has been laying off engineers in the gpu department, dont think they would be doing that if they just wanted patent protection. This lets Intel focus solely on the cpu and just place orders for the gpu which will save them money since R&D is so expensive. But it's still just a rumor until its announced.

If that happens, I'm sure Nvidia will lawyer up.
Intel and AMD working together will gobble up market share. Nvidia will be stuck with discrete solutions.

Also, maybe Nvidia saw this coming and made the push years ago for machine learning ahead of their contract with Intel expiring.
Intel is behind the curve on AI because of their partnership of Nvidia IMO. Nivida has the tech of choice for the future.
 
Here's my read on the situation:

1. As Ryzen information continues to emerge Intel is realizing the entirety of it's x86 markets are at risk of substantial market share erosion, starting with the highest margin and highest profits segments. Intel realizes it's own GPU efforts are hopelessly behind the curve in an increasingly GPU centric world.

2. Intel decided to get on board with AMD's steadily building 'irresitible force' open initiatives and GPU architectures.

3. Part of Intel turning over it's GPU development to the RTG pretty much requires AMD spinning off the RTG. If that happens expect Navi to pop up across the x86, ARM and IBM Power landscapes.

4. With Intel sporting RTG GPU architectures, full integration possible and probable starting in 2019, Nvidia will be pushed out of the consumer and professional compute space.

5. As Ryxen eats it's CPU/APU lunch, Intel will be left with a lot of unused fabrication capacity, and since they would already be engineering RTG GPU architectures into their chips, why not make that unused or under-used capacity available to the RTG, giving further impetus to pushing Nvidia from the space?

6. Intel and AMD could collaborate to create operational duopolies in additional existing and emerging market segments.

7. Lise Su is emerging as perhaps the smartest, savviest and most competent CEO in the world. Raja Koduri as CEO of the RTG wouldn't be far behind.

8. Once Ryzen launches and AMD's market position is secured, it's possible an RTG spinoff will be announced, and BEFORE full disclosure of Intel's agreement with the RTG is announced. To do otherwise would hard crash Intel's stock price and rocket booster AMD's.

Really is a shame the Arctic Ocean ice cover will melt clean away by summer's end throwing the world into an unending drastic and chaotic emergency mode - just as compute tech was REALLY starting to get exciting.
 
Last edited:
7. Lise Su in emerging as perhaps the smartest, savviest and most competent CEO in the world. Raja Koduri as CEO of the RTG won't be far behind.

8. Once Ryzen launches and AMD's market position is secured, it's likely a RTG spinoff will be announced, and BEFORE full disclosure of Intel's agreement with the RTG is announced. To do otherwise would hard crash Intel's stock price and rocket booster AMD's.
.

This is not a smart move from a CEO when one has both a CPU and GPU division, furthermore more costs are spent developing-manufacturing line and plant from the ground up for CPU/APU than Polaris and Vega, so to reduce that market with this deal is a bit insane just for a bit more market on a low margin component GPU orientated for Intel, which means the deal may had come from the GPU division rather than Lisa Su and she cannot control her VPs.
This Intel deal could put Ryzen/APU future at risk, which AMD and investors are expecting to do well as it is needed to 'save' the company.
Cheers
 
Last edited:
This is not a smart move from a CEO when one has both a CPU and GPU division, furthermore more costs are spent developing-manufacturing line and plant from the ground up for CPU/APU than Polaris and Vega, so to reduce that market with this deal is a bit insane just for a bit more market on a low margin component GPU orientated for Intel, which means the deal may had come from the GPU division rather than Lisa Su and she cannot control her VPs.
This Intel deal could put Ryzen/APU future at risk, which AMD and investors are expecting to do well as it is needed to 'save' the company.
Cheers

Nothing substantive in happening at AMD outside of Lisa's knowledge or control. An RTG spinoff would be Intel driven to provide an ironclad legal wall for Intel IP. Everyone else will be making use of AMD's semi-custom unit to integrate AMD or RTG IP with their IP.
 
Nothing substantive in happening at AMD outside of Lisa's knowledge or control. An RTG spinoff would be Intel driven to provide an ironclad legal wall for Intel IP. Everyone else will be making use of AMD's semi-custom unit to integrate AMD or RTG IP with their IP.
Any spinoff would involve the CEOs and their business strategy for their whole business portfolio/divisions meaning it would be incredibly unlikely she would support such a situation, the talk-rumours about splitting originated that this is what Raja wants, not rumours-analysis around Lisa.
Consider how many engineers who are now in the new Radeon division were involved with the CPU-APU before the teams split, that is a serious brain drain for future R&D if she lost the GPU division to Intel.
Let alone crippling one aspect AMD is stronger than Intel at from a business product side (decent APUs and SoCs).
Cheers
 
Last edited:
Here's my read on the situation:

1. As Ryzen information continues to emerge Intel is realizing the entirety of it's x86 markets are at risk of substantial market share erosion, starting with the highest margin and highest profits segments. Intel realizes it's own GPU efforts are hopelessly behind the curve in an increasingly GPU centric world.

2. Intel decided to get on board with AMD's steadily building 'irresitible force' open initiatives and GPU architectures.

3. Part of Intel turning over it's GPU development to the RTG pretty much requires AMD spinning off the RTG. If that happens expect Navi to pop up across the x86, ARM and IBM Power landscapes.

4. With Intel sporting RTG GPU architectures, full integration possible and probable starting in 2019, Nvidia will be pushed out of the consumer and professional compute space.

5. As Ryxen eats it's CPU/APU lunch, Intel will be left with a lot of unused fabrication capacity, and since they would already be engineering RTG GPU architectures into their chips, why not make that unused or under-used capacity available to the RTG, giving further impetus to pushing Nvidia from the space?

6. Intel and AMD could collaborate to create opoerational duopolies in additional existing and emerging market segments.

7. Lise Su is emerging as perhaps the smartest, savviest and most competent CEO in the world. Raja Koduri as CEO of the RTG wouldn't be far behind.

8. Once Ryzen launches and AMD's market position is secured, it's likely an RTG spinoff will be announced, and BEFORE full disclosure of Intel's agreement with the RTG is announced. To do otherwise would hard crash Intel's stock price and rocket booster AMD's.

Really is a shame the Arctic Ocean ice cover will melt clean away by summer's end throwing the world into an unending drastic and chaotic emergency mode - just as compute tech was REALLY starting to get exciting.


Dude. Meet me somewhere private and turn me on to the shit you are smoking. I'll pay good money!!
 
Any spinoff would involve the CEOs and their business strategy for their whole business portfolio/divisions meaning it would be incredibly unlikely she would support such a situation, the talk-rumours about splitting originated that this is what Raja wants, not rumours-analysis around Lisa.
Consider how many engineers who are now in the new Radeon division were involved with the CPU-APU before the teams split, that is a serious brain drain for future R&D if she lost the GPU division to Intel.
Let alone crippling one aspect AMD is stronger than Intel at from a business product side (decent APUs and SoCs).
Cheers

Where did the idea if the RTG was spun off Intel would 'own' it come from?
 
Why would AMD sell its RTG division? it is a money making division that drives a lot of AMD technology.

seems like Intel may want to slash some of its spiraling out of control R&D by outsourcing to another to provide a custom iGPU, while not cheap it is smaller than intels and it would be faster. AMD gets money and royalties, Intel saves costs in iGPU designs and opens die space.
 
Why would anyone buy an AMD APU then? And will 15W parts now become 35W parts? How many actually buys IGP based on graphics performance? And in a way where you dont instantly collapse from laughing while trying to point out the much better and much faster discrete solution.

GCN is anything but power efficient. How had you imagined it should be ARM competitor worth against graphics designed for low power? Hell, AMD cant even compete on something as simple as quicksync and decode formats.
:rolleyes:
:LOL:
Would it matter to AMD if they sell APU's if all APU's have AMD in it? That is X86/64 ones.

Plus AMD could sell more discreet cards if the Intel CPU can better work together with it.

I agree previous versions of GCN are not as power efficient as like Nvidia - that does not indicate that they cannot be with Intel R&D, AMD advancements. If AMD can cut down geometry processing by many folds, only processing what will be seen (increase efficiency = less power) load only items needed in memory (increase efficiency). Plus it is no longer GCN but a NCU core :p.

Now think about this, could not Intel own 30% or 40% of RTG while AMD owns the rest or it is further split up? Tying resources of more then one company?

You know with the down turn of PC's in general for numbers sold or flat line - Intel could open up their fab shops more to like AMD to make more of a profit. In other words Intel has much to gain as well as AMD if they join up or work together.
 
:rolleyes:
:LOL:
Would it matter to AMD if they sell APU's if all APU's have AMD in it? That is X86/64 ones.

Plus AMD could sell more discreet cards if the Intel CPU can better work together with it.

I agree previous versions of GCN are not as power efficient as like Nvidia - that does not indicate that they cannot be with Intel R&D, AMD advancements. If AMD can cut down geometry processing by many folds, only processing what will be seen (increase efficiency = less power) load only items needed in memory (increase efficiency). Plus it is no longer GCN but a NCU core :p.

Now think about this, could not Intel own 30% or 40% of RTG while AMD owns the rest or it is further split up? Tying resources of more then one company?

You know with the down turn of PC's in general for numbers sold or flat line - Intel could open up their fab shops more to like AMD to make more of a profit. In other words Intel has much to gain as well as AMD if they join up or work together.

RTG is part of AMD and for that a major resolution would have to be passed for a selling off, that would also require a shareholders approval, the likelihood of AMD selling anything off to Intel is about as likely as pigs that fly in the middle of July. AMD is not selling off, Intel came to them as a private client to do custom GPU's, AMD renders a service for cost and because it will Brand AMD they will get royalties per part. Given Intels market share that is substantial revenue.

Does it affect APU's? maybe but then again like all power users iGPU is irrelevant, and if AMD release a Zen+Vega and HBM APU SoC, then I am sure it will sell like candy to a fat kid in a candy store.
 
:rolleyes:
:LOL:
Would it matter to AMD if they sell APU's if all APU's have AMD in it? That is X86/64 ones.

Plus AMD could sell more discreet cards if the Intel CPU can better work together with it.

I agree previous versions of GCN are not as power efficient as like Nvidia - that does not indicate that they cannot be with Intel R&D, AMD advancements. If AMD can cut down geometry processing by many folds, only processing what will be seen (increase efficiency = less power) load only items needed in memory (increase efficiency). Plus it is no longer GCN but a NCU core :p.

Now think about this, could not Intel own 30% or 40% of RTG while AMD owns the rest or it is further split up? Tying resources of more then one company?

You know with the down turn of PC's in general for numbers sold or flat line - Intel could open up their fab shops more to like AMD to make more of a profit. In other words Intel has much to gain as well as AMD if they join up or work together.

What do you think revenue and margins is on a theoretical MCM GPU part is compared to selling an APU?
 
From what Kyle said previously, Intel only needs patent protection from AMD, not iGPUs.
Intel current patent deal with Nvidia didn't put a single Nvidia iGPU inside an Intel CPU.

The supposed deal is still not proven. So as such it doesn't exist (yet).

Intel is also hiring a lot of GPU people, so that excludes the wild fanboy dreams of AMD GPUs everywhere. Not that it even made sense to begin with outside the most wild and crazy unicorn and fairy tale dreams.

And as you mention, Intel didn't put an Nvidia GPU into its part, despite an Nvidia GPU is superior to AMD GPUs in all metrics.
 
What do you think revenue and margins is on a theoretical MCM GPU part is compared to selling an APU?
Well considering AMD will not have to have to make it, store it, market it, distribute it, warranty it - answer is it depends upon how much AMD can get per APU or MCM in this case - so variable and no set answer. Plus if AMD has a marketable APU that can sell at a price that even makes a profit - they could actually lose money with an APU - wait they have already.

Now I hope you know all of this, most of the thread is conjecture, brain storming and mostly wild ass guesses. In other words I wouldn't take it too serious until any real deals are done.
 
The supposed deal is still not proven. So as such it doesn't exist (yet).

Intel is also hiring a lot of GPU people, so that excludes the wild fanboy dreams of AMD GPUs everywhere. Not that it even made sense to begin with outside the most wild and crazy unicorn and fairy tale dreams.

And as you mention, Intel didn't put an Nvidia GPU into its part, despite an Nvidia GPU is superior to AMD GPUs in all metrics.
So black and white you be ;). All metrics except for unicorns and fairy dust :cool:.
 
RTG is part of AMD and for that a major resolution would have to be passed for a selling off, that would also require a shareholders approval, the likelihood of AMD selling anything off to Intel is about as likely as pigs that fly in the middle of July. AMD is not selling off, Intel came to them as a private client to do custom GPU's, AMD renders a service for cost and because it will Brand AMD they will get royalties per part. Given Intels market share that is substantial revenue.

Does it affect APU's? maybe but then again like all power users iGPU is irrelevant, and if AMD release a Zen+Vega and HBM APU SoC, then I am sure it will sell like candy to a fat kid in a candy store.
AMD - first enthusiast APU :sleep:. Custom designs which AMD is doing well in (PS4/pro, Xbox 360/1/. . .) I agree is probably most likely scenario as well since RTG proven to be able to keep designs separate from competing companies (Sony and Microsoft). Below is somewhat old article at Forbes but makes sense.

http://www.forbes.com/sites/tiriasr...ensing-radeon-graphics-to-intel/#292c41a37719
 
I highly doubt this licensing deal will be any more lucrative than those console deals.

One possibility I can think of is that Intel is trying to squeeze Nvidia, to try and regain some marketshare in the HPC space.

Another possibility is that this is a test solution, and a more complex form of it consisting of Xeon Phi + AMD GPU co-processor will pop up in the future. You then get the best of both HPC worlds, I guess?
 
I highly doubt this licensing deal will be any more lucrative than those console deals.

One possibility I can think of is that Intel is trying to squeeze Nvidia, to try and regain some marketshare in the HPC space.

Another possibility is that this is a test solution, and a more complex form of it consisting of Xeon Phi + AMD GPU co-processor will pop up in the future. You then get the best of both HPC worlds, I guess?
They both want to squeeze Nvidia, this deal assists AMD from that perspective but allows Intel to broaden their market.

I doubt this effects HPC space, Intel is ploughing a lot of resources into both Altera and critically Nervana (this is more important than anything AMD could provide for them in this context).
Nervana is going to be software+hardware, it will be its own hardware solution and also integrated into the next gen Xeon Phi's.
The HPC 'arms tech race' is really just taking off between Intel and Nvidia, and in some of the other segments other specialist-niche companies against both of them.
Cheers
 
What do you think revenue and margins is on a theoretical MCM GPU part is compared to selling an APU?

Think of it this way: How much money does AMD make off an Intel chip now? *none*

Sure they would love to sell their own products but since they don't own their own fabs they already have a lot of overhead so profit in the budget segment is never going to be too lucrative.

Having their IP in Intel products at least guarantees some revenue when, not if, Intel has better CPU sales then Zen.

Even if AMD pulls a miracle and utterly stomps Intel in a couple of weeks it wont last more than a year two before Intel surpasses them again. Best case is they can manage to trade blows for years like they did before.

On top of all that AMD could sell the best thing ever and Intel could release the second coming of Prescott and the market shares would barely budge
 
Think of it this way: How much money does AMD make off an Intel chip now? *none*

Sure they would love to sell their own products but since they don't own their own fabs they already have a lot of overhead so profit in the budget segment is never going to be too lucrative.

Having their IP in Intel products at least guarantees some revenue when, not if, Intel has better CPU sales then Zen.

Even if AMD pulls a miracle and utterly stomps Intel in a couple of weeks it wont last more than a year two before Intel surpasses them again. Best case is they can manage to trade blows for years like they did before.

On top of all that AMD could sell the best thing ever and Intel could release the second coming of Prescott and the market shares would barely budge

This

Intel will sell a lot, AMD will have extra revenue, irrespective of APU's
 
Think of it this way: How much money does AMD make off an Intel chip now? *none*

Sure they would love to sell their own products but since they don't own their own fabs they already have a lot of overhead so profit in the budget segment is never going to be too lucrative.

Having their IP in Intel products at least guarantees some revenue when, not if, Intel has better CPU sales then Zen.

Even if AMD pulls a miracle and utterly stomps Intel in a couple of weeks it wont last more than a year two before Intel surpasses them again. Best case is they can manage to trade blows for years like they did before.

On top of all that AMD could sell the best thing ever and Intel could release the second coming of Prescott and the market shares would barely budge
And then they start to take losses on their CPU/APU/SoC business.
The positives some put forward regarding Intel deal cannot have it both ways, and this is compounded that the CPU/APU division has been eating a serious amount of the R&D budget (much more than the dGPU) and investors-analysts see this as what will push the company forward, in other words AMD needs CPU/APU/SoCs to succeed.
Cheers
 
And then they start to take losses on their CPU/APU/SoC business.
The positives some put forward regarding Intel deal cannot have it both ways, and this is compounded that the CPU/APU division has been eating a serious amount of the R&D budget (much more than the dGPU) and investors-analysts see this as what will push the company forward, in other words AMD needs CPU/APU/SoCs to succeed.
Cheers

They already take losses on their CPU/APU division. If they rely on APU income and it never shows up they will be out of business in a few years. On the other hand monetizing their graphics IP can fund CPU R&D to keep them competitive down the road.

Their definition of success is not you being able to buy an awesome AMD CPU. Success is having a job tomorrow via providing ROI to their investors and if at the end of the day they simply cant compete in the CPU market then they won't.

I hope it doesn't come to that but I don't blame them for not placing all their bets on Zen, no matter how great it may end up being.
 
And then they start to take losses on their CPU/APU/SoC business.
The positives some put forward regarding Intel deal cannot have it both ways, and this is compounded that the CPU/APU division has been eating a serious amount of the R&D budget (much more than the dGPU) and investors-analysts see this as what will push the company forward, in other words AMD needs CPU/APU/SoCs to succeed.
Cheers
Eating R&D, but producing components that get shared among all products. Infinity fabric for instance is used by CPU, GPU, and APU. Execution units might be shared. The Linley paper for instance laid out some interesting changes to FMAs. Those changes could be more practical with packed math and potentially used with Vega. It would be the equivalent of hyperthreading applied to a GPU with an op cache essentially providing additional operands. All the positioning AMD appears to be doing is related to making components easily integrated together into possible custom designs.

dGPU, especially at the bottom of the market, is likely going away with a wider move towards SFF. More appropriately, dGPU is transitioning towards MCM for SFF. Some degree of configurability is maintained, better power/perf from proximity, and the obvious smaller form factor being the primary driver. A model where RTG for example is making custom designs, including for desktop GPU, might be practical. Intel gets a custom design, Apple a custom design, board partner custom design. Removing AMD from the actual manufacturing side of things.
 
They already take losses on their CPU/APU division. If they rely on APU income and it never shows up they will be out of business in a few years. On the other hand monetizing their graphics IP can fund CPU R&D to keep them competitive down the road.

Their definition of success is not you being able to buy an awesome AMD CPU. Success is having a job tomorrow via providing ROI to their investors and if at the end of the day they simply cant compete in the CPU market then they won't.

I hope it doesn't come to that but I don't blame them for not placing all their bets on Zen, no matter how great it may end up being.
Is is also APU/SoC not just the CPU.
This deal potentially eats into the revenue potential of this heavy R&D project.
It does not need much more funding as they are looking to keep this design going for 8 years with a mild evolution 3-4 years from now, but it importantly needs to be monetized.

I got a feeling this topic is going to split the community until we get to see the details and broadness of any deal that may be happening.
Cheers
 
Last edited:
Really is a shame the Arctic Ocean ice cover will melt clean away by summer's end throwing the world into an unending drastic and chaotic emergency mode - just as compute tech was REALLY starting to get exciting.

what the?
 
Mercedes supply race engines to williams and force india racing, do they threaten Mercedes? Nope not at all. Any spec build by amd will be for a certain spec eg: intel would not spend 4bn ordering a Vega 10 spec igpu with HBM for pentiums and i3's. It would likely only be for high end if assumed the case arose. I honestly think Intel wants to remove dead weight off their die and save costs on a petty design while offering notably better solutions across the range eg: canning integrated solutions to outsource for generic spec. This will cut cists for intel, give amd cashflow while AMD reserves high end integrated solutions to its APU line.
 
Think of it this way: How much money does AMD make off an Intel chip now? *none*

Sure they would love to sell their own products but since they don't own their own fabs they already have a lot of overhead so profit in the budget segment is never going to be too lucrative.

Having their IP in Intel products at least guarantees some revenue when, not if, Intel has better CPU sales then Zen.

Even if AMD pulls a miracle and utterly stomps Intel in a couple of weeks it wont last more than a year two before Intel surpasses them again. Best case is they can manage to trade blows for years like they did before.

On top of all that AMD could sell the best thing ever and Intel could release the second coming of Prescott and the market shares would barely budge

You avoided the question because it completely invalidates your post.

It would hypothetically put AMDs CPU division in the grave. You see the conflicts dont you.
 
Mercedes supply race engines to williams and force india racing, do they threaten Mercedes? Nope not at all. Any spec build by amd will be for a certain spec eg: intel would not spend 4bn ordering a Vega 10 spec igpu with HBM for pentiums and i3's. It would likely only be for high end if assumed the case arose. I honestly think Intel wants to remove dead weight off their die and save costs on a petty design while offering notably better solutions across the range eg: canning integrated solutions to outsource for generic spec. This will cut cists for intel, give amd cashflow while AMD reserves high end integrated solutions to its APU line.

How many CPUs sell due to the IGP alone and not either CPU performance, perf/watt or cost?

Dead weight? Please explain. I cant wait to get a laugh out of that one. You better look on APU first and then remember to check mobile graphics benchmarks, they are far from desktop ;)

And why would you make deals with AMD, if you dont get what you want?
 
How many CPUs sell due to the IGP alone and not either CPU performance, perf/watt or cost?

Dead weight? Please explain. I cant wait to get a laugh out of that one. You better look on APU first and then remember to check mobile graphics benchmarks, they are far from desktop ;)

And why would you make deals with AMD, if you dont get what you want?

I did state outside the top end parts the iGPU still has a market because believe it or not, not everyone needs a 1080.

As per Chris Angelini's review on Skylakes iGPU the AMD APU's still do well despite severely bottlenecking the iGPU component. We will wait on Raven Ridge APU's to better determine performance.

Intel would get what they want, a cheaper than developing themselves iGPU and cross patent protection. They can dump a entry level custom built iGPU save die space and put it towards X86
 
I did state outside the top end parts the iGPU still has a market because believe it or not, not everyone needs a 1080.

As per Chris Angelini's review on Skylakes iGPU the AMD APU's still do well despite severely bottlenecking the iGPU component. We will wait on Raven Ridge APU's to better determine performance.

Intel would get what they want, a cheaper than developing themselves iGPU and cross patent protection. They can dump a entry level custom built iGPU save die space and put it towards X86

AMDs APUs are outright a disaster in mobile space because they cant reach the TDP needed and throttle like mad. And GCN is a big party of that problem.

There isn't any saved die space. And it needs to be integrated, not MCM. And save what? How much would this cost? Socket size? Power delivery? There are so many lose ends in this fairy tale adventure that its a tragic comic entertainment.

And again, Nvidia can do everything better. Why go with the subpair provider if they was to ditch their own IGP line? Why would they even remotely want to pick Polaris/Vega over Pascal/Volta? You know Nvidia need less transistors, less memory bandwidth, less power for the same performance and they got much better video decode and feature support. Not having to use 40W to play a bluray movie because AMD is trying to solve a fundamental codec support issue with OpenCL that leaves weak cards and APUs out for the higher tier codecs.
 
Last edited:
AMDs APUs are outright a disaster in mobile space because they cant reach the TDP needed and throttle like mad. And GCN is a big party of that problem.

There isn't any saved die space. And it needs to be integrated, not MCM. And save what? How much would this cost? Socket size? Power delivery? There are so many lose ends in this fairy tale adventure that its a tragic comic entertainment.

And again, Nvidia can do everything better. Why go with the subpair provider if they was to ditch their own IGP line? Why would they even remotely want to pick Polaris/Vega over Pascal/Volta? You know Nvidia need less transistors, less memory bandwidth, less power for the same performance and they got much better video decode and feature support. Not having to use 40W to play a bluray movie because AMD is trying to solve a fundamental codec support issue with OpenCL that leaves weak cards and APUs out for the higher tier codecs.

nobody was mentioning mobile, the point is moot.

Vega die space is less than intels iGPU solutions.

Are Intel partnering with Nvidia? I know you have a love affair with everything Intel and Nvidia but your beloved Nvidia is actually wanting to put Intel out of business targeting areas of computing that Intel have a foothold in. Why would Nvidia want to even contemplate helping Intel given the bad blood both share. Right now Nvidia has zero incentive to help Intel and probably wish that no graphics licence is past by AMD to Intel so they can sue Intel for love and chappies. Wait until Nvidia release competition to Xeon Phi, then the bad blood will start, and Nvidia are not going to be a push over.

Intel AMD can mutually exist without major ramifications on Intel's spreadsheets, AMD pocket some money but ultimately AMD is more orientated to specific market spaces that doesn't infringe on Intel's capital.
 
nobody was mentioning mobile, the point is moot.

It has to fit mobile too doesn't it. You know the biggest consumer market.

Vega die space is less than intels iGPU solutions.

Documentation please.

Are Intel partnering with Nvidia? I know you have a love affair with everything Intel and Nvidia but your beloved Nvidia is actually wanting to put Intel out of business targeting areas of computing that Intel have a foothold in. Why would Nvidia want to even contemplate helping Intel given the bad blood both share. Right now Nvidia has zero incentive to help Intel and probably wish that no graphics licence is past by AMD to Intel so they can sue Intel for love and chappies. Wait until Nvidia release competition to Xeon Phi, then the bad blood will start, and Nvidia are not going to be a push over.

Intel AMD can mutually exist without major ramifications on Intel's spreadsheets, AMD pocket some money but ultimately AMD is more orientated to specific market spaces that doesn't infringe on Intel's capital.

All the same applies to AMD despite you obvious "love affair". But unlike AMD, Nvidia isn't selling competing x86 CPUs with IGP.
 
Back
Top