pendragon1
Extremely [H]
- Joined
- Oct 7, 2000
- Messages
- 52,196
that vid is 1 1/2 old. so "went looking for/had saved for future use" is more appropriate. I'm sure you ranted about it back then too...
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
From ATI to AMD back to ATI? A Journey in Futility @ [H]
- Thread starter FrgMstr
- Start date
Until AMD stops blatantly lying like this in their marketing I won't even consider looking at them.
The deal seems like a good one for the price, but AMD haven't yet released a product that they haven't wildly lied about the performance of in at least 5 years, so the above statement applies. I won't support a company that consistently and so obviously lies about their own products.
I haven't seen AMD lying about Polaris, they kept expectations low. If people want to believe the hype generated by wccftech and shit then its their fault. AMD even tried to shut him by getting under NDA about his bullshit stories about the card overclocking to 1600mhz.
So they have improved a bit since then, it was a bad decision to go with HBM for them. It was probably because they had a part in developing it so they had vested interest in it. But I do believe 8GB HBM2 should be pretty decent amount of memory. The bandwidth of HBM does help it a bit no doubt. But no 4gb hbm is not better than 12gb gddr5. But I wont argue 4GB HBM has held surprisingly well on the fury cards.
N4CR
Supreme [H]ardness
- Joined
- Oct 17, 2011
- Messages
- 4,947
Guess you were under a rock when they launched the 7970? 290x did really damn well too once freed from the reference coolers, I remember it upsetting a certain card called a titan.
I have no problems is AMD and intel work toghether in video and both benefit, that would be fanstastic.. Imagine Intel-branted and built AMD video cards. As long as they keep competing in CPU, should be fine.
Then again this could be a game to get to nvidia to play ball.
Then again Intel could be making moves to suck in AMD with out much regulatory intervention (not that there would be much to say with ARM around anyway)
Then again this could be a game to get to nvidia to play ball.
Then again Intel could be making moves to suck in AMD with out much regulatory intervention (not that there would be much to say with ARM around anyway)
Shintai
Supreme [H]ardness
- Joined
- Jul 1, 2016
- Messages
- 5,678
the 2011 settlement is more about the issue with nvidia making chipset for intel motherboard and not about intel infringing nvidia gpu patent. in 2004 both company already made cross licensing deal which allow intel to continue developing their gpu without infringing nvidia patent while at the same time nvidia was allowed to make chipsets for intel motherboard.
the issue starts when intel said nvidia does not have the right to make chipset for intel mobo starting from certain generation.
so the settlement back then has nothing to do with intel infringing nvidia patent. they already make cross licensing deal back in 2004. the agreement did expire in 2011 but with everything that's going somehow the settlement also include to renew the agreement made in 2004 because without those access intel will definitely got into another trouble with nvidia.
source: http://www.anandtech.com/show/4122/intel-settles-with-nvidia-more-money-fewer-problems-no-x86
InquisitorDavid
Weaksauce
- Joined
- Jun 27, 2016
- Messages
- 90
Wow, my brain died a little bit when he started saying 'system memory and swap things around'.
It's only awesome when AMD does it.™
Shintai
Supreme [H]ardness
- Joined
- Jul 1, 2016
- Messages
- 5,678
Vega is now "Middle of 2017" btw in case anyone missed it.
http://seekingalpha.com/article/402...ent-presents-barclays-global-technology-media
http://seekingalpha.com/article/402...ent-presents-barclays-global-technology-media
- Joined
- May 18, 1997
- Messages
- 55,634
Interesting...
Question"
Okay. On the licensing front, I won’t put you on the spot about specific recent rumors but say at least there has been some talk that you might have a partnership with one of your largest competitors, rather than talking about that specifically, could you just talk about your appetite more broadly for other licensing partnerships and what type of willingness you’d have to do those?
Amswer:
Yes, I think, we've talked about IP monetization in particular partnerships. So, AMD with the treasure full of IP that we have, we have 10,000 plus patterns, half of them are US-based. And we saw over the last two or three years that we can go ahead and partner with folks, where we don't want to directly enter market. We will be very careful in terms of who we partner with, where it doesn't come back and directly compete with us in areas that we want to go put products in, and very happy to do on a fair basis a deal with the partner to go ahead and monetize that IP, get cash, benefit the P&L and balance sheet in particular.
WTF, if your are a journalist would you not put them on the spot? They going to ban you from their events?
- Joined
- May 18, 1997
- Messages
- 55,634
Interesting quotes, I would suggest you register and read the whole thing.
http://seekingalpha.com/article/402...ent-presents-barclays-global-technology-media
Alibaba is introducing a current generation of GPUs, but we do have Vega products coming out in the middle part of 2017. And that should help us from an overall standpoint of being more competitive in a space that, for example, our competitor has taken that world by storm.
______________-
And we’re well-positioned to do that. And we’re the only one company that can do in both in the CPU side and the GPU side. And by the way, the products are getting competitive, both of them, the investment we’ve had over the last couple of years, they are both competitive products being launched in 2017, CPU on a Zen side of it and GPU with the Vega products that’s coming up.
Going to 2017, the Zen Core that we've been working on for multiple number of years, clean chip design, the first product which we call Summit Ridge, targeted for the high-end desktops, we’ll launch in Q1 of 2017. And that is really addressing an area, which is a $4 billion market that we don't really compete in today, and we're looking forward to that product being the first proof point, which a lot of people want to know is Zen on track, i.e., have their customers or their shipments.
____________
Unidentified Company Representative
Do you think, are you pretty confident you can get back to, I guess, prior share levels like what we might have seen five or six years ago?
Devinder Kumar
I think, it is safe to say that the competitive product that we are introducing, we can improve our revenue share over the next multiple number of years, because they are markets, like I said, the high-end desktop market is $4 billion . Today, we don't compete in that. And any share that we get from there is accretive from a revenue share standpoint, is accretive from a gross margin standpoint, and accreting from the gross margin dollars that are generated to go ahead and fund the future investments from an R&D standpoint.
The other thing that I should mention is, one of the one of the concerns when you do a multi-year arrangement from a product roadmap standpoint is, hey, the processor technology is going to be there, is the processor technology going to be available. Today, we’re in the FinFET technology known as 14 and 16 and later we are going to go to 7.
________________
Unidentified Company Representative
So beyond, obviously, the server market is dominated by x86 today, but there has been a lot of high excitement around the possibility of ARM-based products, certainly you played as well, QUALCOMM just announced a 10 nanometer product. Can you just talk a little bit about your expectations near-term for how ARM-based server products will impact the market, and what type of opportunities you see for AMD around that?
Devinder Kumar
Yes. So if you look back just the last couple of years from an AMD standpoint, the one thing a lot of people asked questions earlier today when I met some folks, what changed over the last couple of years? I think it’s all about focus and execution. But more importantly, we had to make decisions about where we want to invest. And you’re going to ask me probably about the GPUs, the accelerator space, the architecture intelligent, and that's an area we choose to invest in, because we see huge potential in a growing market, where there are two players and we can go capture market share.
But we also have to make decisions about where not to invest. And, frankly, ARM was one of these areas we said in the TAM in the x86 space from a data center standpoint and the gaming space and the immersive platform that we have is $15 million, and we’re sub-10% today from our overall market share.
We can go invest and either take market share, or gain market share in the growing market is the area to go invest in. We do have ARM products out there. In fact, in the semi-custom space, we asked a lot about the three design wins we talked about last year, a couple of them already known publicly the Sony PlayStation Pro, the project scope of Microsoft that’s coming in the second-half of 2017.
___________________
Devinder Kumar
Time will tell. I think, we’re more about the open platform, NVIDIA has CUDA. But there are ways in which you can take the CUDA core and go ahead and port that over to AMD GPUs and we’ll see how our customers want to do that. The customers are definitely interested just like in any particular space when you have competitive products, the customers want choice.
___________________
Devinder Kumar
I think overall GPU has been a growth market because of the new areas that GPUs are going to be, it’s definitely a growth market, definitely a growth market. Discrete GPU for the PC space, our market share is low. We've made some strides, and I think, we continue to gain market share. But really GPU is going to become more than just about GPUs that going to the PC space.
GPUs is kind of the new trend. You can say it’s on fire in terms of everything that you're reading about where the markets could be. And I think that's where the future potential is from the longer-term for AMD. We've always been going to the CPU house. And I think just based on the things that we have done over the last year or so, in particular, the data center space that we just talked about, GPU is definitely an area of opportunity and growth in a growing market with only two competitors in that area, that's pretty exciting.
___________________________
Devinder Kumar
I would say, this is an area where we are very careful about what we say. But I would say, the pipeline is healthy. There are opportunities, and in particular, even the Zen-based products, I think, the emphasis is going to be even higher that can we now take, we with our partners take the Zen-based core into a semi-custom business that is multiple millions of units, and I think that's a possibility.
http://seekingalpha.com/article/402...ent-presents-barclays-global-technology-media
Alibaba is introducing a current generation of GPUs, but we do have Vega products coming out in the middle part of 2017. And that should help us from an overall standpoint of being more competitive in a space that, for example, our competitor has taken that world by storm.
______________-
And we’re well-positioned to do that. And we’re the only one company that can do in both in the CPU side and the GPU side. And by the way, the products are getting competitive, both of them, the investment we’ve had over the last couple of years, they are both competitive products being launched in 2017, CPU on a Zen side of it and GPU with the Vega products that’s coming up.
Going to 2017, the Zen Core that we've been working on for multiple number of years, clean chip design, the first product which we call Summit Ridge, targeted for the high-end desktops, we’ll launch in Q1 of 2017. And that is really addressing an area, which is a $4 billion market that we don't really compete in today, and we're looking forward to that product being the first proof point, which a lot of people want to know is Zen on track, i.e., have their customers or their shipments.
____________
Unidentified Company Representative
Do you think, are you pretty confident you can get back to, I guess, prior share levels like what we might have seen five or six years ago?
Devinder Kumar
I think, it is safe to say that the competitive product that we are introducing, we can improve our revenue share over the next multiple number of years, because they are markets, like I said, the high-end desktop market is $4 billion . Today, we don't compete in that. And any share that we get from there is accretive from a revenue share standpoint, is accretive from a gross margin standpoint, and accreting from the gross margin dollars that are generated to go ahead and fund the future investments from an R&D standpoint.
The other thing that I should mention is, one of the one of the concerns when you do a multi-year arrangement from a product roadmap standpoint is, hey, the processor technology is going to be there, is the processor technology going to be available. Today, we’re in the FinFET technology known as 14 and 16 and later we are going to go to 7.
________________
Unidentified Company Representative
So beyond, obviously, the server market is dominated by x86 today, but there has been a lot of high excitement around the possibility of ARM-based products, certainly you played as well, QUALCOMM just announced a 10 nanometer product. Can you just talk a little bit about your expectations near-term for how ARM-based server products will impact the market, and what type of opportunities you see for AMD around that?
Devinder Kumar
Yes. So if you look back just the last couple of years from an AMD standpoint, the one thing a lot of people asked questions earlier today when I met some folks, what changed over the last couple of years? I think it’s all about focus and execution. But more importantly, we had to make decisions about where we want to invest. And you’re going to ask me probably about the GPUs, the accelerator space, the architecture intelligent, and that's an area we choose to invest in, because we see huge potential in a growing market, where there are two players and we can go capture market share.
But we also have to make decisions about where not to invest. And, frankly, ARM was one of these areas we said in the TAM in the x86 space from a data center standpoint and the gaming space and the immersive platform that we have is $15 million, and we’re sub-10% today from our overall market share.
We can go invest and either take market share, or gain market share in the growing market is the area to go invest in. We do have ARM products out there. In fact, in the semi-custom space, we asked a lot about the three design wins we talked about last year, a couple of them already known publicly the Sony PlayStation Pro, the project scope of Microsoft that’s coming in the second-half of 2017.
___________________
Devinder Kumar
Time will tell. I think, we’re more about the open platform, NVIDIA has CUDA. But there are ways in which you can take the CUDA core and go ahead and port that over to AMD GPUs and we’ll see how our customers want to do that. The customers are definitely interested just like in any particular space when you have competitive products, the customers want choice.
___________________
Devinder Kumar
I think overall GPU has been a growth market because of the new areas that GPUs are going to be, it’s definitely a growth market, definitely a growth market. Discrete GPU for the PC space, our market share is low. We've made some strides, and I think, we continue to gain market share. But really GPU is going to become more than just about GPUs that going to the PC space.
GPUs is kind of the new trend. You can say it’s on fire in terms of everything that you're reading about where the markets could be. And I think that's where the future potential is from the longer-term for AMD. We've always been going to the CPU house. And I think just based on the things that we have done over the last year or so, in particular, the data center space that we just talked about, GPU is definitely an area of opportunity and growth in a growing market with only two competitors in that area, that's pretty exciting.
___________________________
Devinder Kumar
I would say, this is an area where we are very careful about what we say. But I would say, the pipeline is healthy. There are opportunities, and in particular, even the Zen-based products, I think, the emphasis is going to be even higher that can we now take, we with our partners take the Zen-based core into a semi-custom business that is multiple millions of units, and I think that's a possibility.
Shintai
Supreme [H]ardness
- Joined
- Jul 1, 2016
- Messages
- 5,678
It all sounds so familiar doesn't it. Oh right, previous releases 
Multiple years to reach 10% = pipedream.
But its the classic tapdance show. Also funny when he talks positive about GPU, yet in reality AMD cuts R&D massively in favour of CPU. (See SEC fillings.)
Multiple years to reach 10% = pipedream.
But its the classic tapdance show. Also funny when he talks positive about GPU, yet in reality AMD cuts R&D massively in favour of CPU. (See SEC fillings.)
Last edited:
So I guess that means they will launch Vega mid-late Q1, with availability late Q2; there could be logistical semantics as they may be releasing it to core clients earlier (or later depending how it goes) and then whether batches 1st to prosumer/HPC or enthusias PC gaming market.
Does seem to be making the dual Polaris more and more likely as a launch linked to current AoTS performance 'leak', unless it is a larger die Polaris (no-one is expecting that though).
Cheers
Does seem to be making the dual Polaris more and more likely as a launch linked to current AoTS performance 'leak', unless it is a larger die Polaris (no-one is expecting that though).
Cheers
Last edited:
LOL no wonder why people thought HBM took care of high resolution gaming, man that is gotta be the best freakin made up shit in the world and then people fell for it? What they forgot the pci-e bus has always been the bottleneck when transferring over from system memory?
Shintai
Supreme [H]ardness
- Joined
- Jul 1, 2016
- Messages
- 5,678
No,no, I was told multiple times by very dedicated forum people that HBM isn't like GDDR. 4GB isn't 4GB! And using crossfire gives you double the VRAM instead of a mirror.
Its amazing how much crap desperate people gobble up and how much crap that guy can feed to the gossip hungry mob.
shad0w4life
Gawd
- Joined
- Jun 30, 2008
- Messages
- 689
Not sure why nvidia gets a free pass, 1070 has been assigned to CSGO idling/NHL streaming. I've had enough of it and gsync garbage I'd rather play new games on xbone & ps4, and my older dx9/10 games on 7870. At least they will run properly.
Presbytier
[H]ard|Gawd
- Joined
- Jun 21, 2016
- Messages
- 1,058
I run a 1080 which is just a beefier 1070 and experience absolutely no problems. The thing people are commenting on is the bullshit marketing hype AMD throws out there for very lackluster products. It would be like Ford releasing a new pinto and trying to get everyone to believe it is a corvette killer.
Ugh. Amd always said for months that rx480 will be mainstream card. Starting to beat a dead horse again. Did they want it to be little faster I am pretty sure they did. But the price also reflects that as well.
Hype is purely created by sites like wccftech and some dude typing bullshit in a forum lol. AMD never really said anything about it being a overclocking monster this time. They actually tried to shut it up. By offering wccftech a card to get him up under NDA and shut up.
I think he's referring to Vega products for the data center. There isn't enough information there to conclude that consumer GPUs based on Vega won't come out earlier.
IIRC, Vega for enterprise was slated first, consumer second.
Their agreement with Nvidia expires in March 2017. So they have to sign a new deal with Nvidia if they want to continue selling new printings of their old designs.
Or sign with the other company AMD... and use a new design.
Yes signing with AMD would cover off some overlapping patents that AMD and Nvidia aren't willing to sue each other over. Still their design would likely have to change quite a bit if they change to AMD. I do believe that is what they are going to do because Intel hates NV with a passion... perhaps more then AMD right now. Radeon has name recognition that intels own products don't.
So come march they either need to resign with Nvidia likely at higher $ then the previous deal. Or sign a deal with AMD for IP and.... stop fabbing chips with their current (nvidia cross licence)GPU and begin fabbing chips with a new (AMD cross licence) GPU. That means they would have to design a whole new Intel GPU that skirts the Nvidia patents by using the AMD ones instead.
This is why its more likely they would simply take a AMD GPU design and slot it into their own die. Why re invent the wheel. When you consider how hard up AMD is for cash... the savings Intel would be looking at would likely be pretty sizable over redesigning a GPU, or even continuing to pay Nvidia. They also get chips their marketing dept can run with.
Last edited:
You probably have the HP laptop which for which it's known that GSync doesn't work on it. Blame HP for making a shitty laptop; blame yourself for not researching before you bought it.
Are you sure about that? Do you have a link?
That video probably did more damage to AMD's credibility, than the bull dozer... Nvidia has been using cuda cores asymmetrically since the 200 series replaced the 8800 GTX. So knowing that the Xbox One uses a ring bus, designed off the old token ring Ethernet that used BNC and Vampire connectors... that might be why he through it had to be done in order... That might be why the Raedon group got to do their own thing. You can buy a 32 Gigabyte size card right now (assuming they are not sold out) from AMD as fire pro videocard. That kinda flies in the face of everything he said. The Xbox One code when it comes over to the PC was a mess at first because the code designed for the ring bus makes a bunch of calls that don't actually do anything. That was why Microsoft moved to Direct X 12 for windows ten. Most of the drivers had to be rewritten and almost everything hit several layers of vitalization before it hits the hard ware to prevent bad drivers from breaking hardware. It add latency but stops some of the worst problems of a person looking at smoking hardware. Some of the open standards get a lot closer to metal but don't have the fake optimizations, well mostly.
Having a larger memory set to work off means that if you know the spots are open you simply write it instead or moving or erasing what is there. Anytime you have to move over the system bus as slow as it is, you get hit with far greater latency because traveling down a wire the width of your thumb verse traveling several inches are magnitudes slower. It will still be fast perceptively much like plugging in a fiber optic cable and seeing the light on a wall next to you as the light travels down the cable... but computers are simply really fast abacuses. They can move data into and out of registers be it memory or processing units, but they are still only adding and subtracting very fast at their most basic function.
I did not see anything about the deal between Intel and AMD, but I did get to attend a private function where AMD was the main sponsor, as the newly minted Raedon group was trying to help Autodesk sell their products and sell more Fire Pro cards, as well as get people looking at where the Raedon Group wanted to be 12 to 18 months from now. They were targeting people who influence and use the workstation cards to for Engineering, Artistic Visualization, Disciples that merge those two like Architecture and Video games, and the simulation and compute crowds. It was during the compute conference so it was not surprising that they wanted to show off to people who might have to plan budgets for their companies two to four years out. So likely that is also driving the stock as they showed off where the Raedon group is aiming at. I made a joke about the larger memory addressing but until I heard that they had publicly announced it some where I was not sure how much they wanted people keeping quiet about it. Direct X is the major API people use, with open gl having some impact as there is less over head on linux machines, by default since you only load the modules you need to do the tasks you want to do.
If I say unloaded the email and notification system on windows, windows closes the gui and restarts, on linux their are a handful of systems that have to run at all times and gui is not even one of them. So if I am rendering from the command line I can use all that lost over head to render faster. If I want to see what what I am rendering I need to render on a machine with enough of the OS running to either look at the rendered output or rendering in the screen. Many people run windows because it is easier. You install the IDE and the installers create all the dependent parts. My guess is Intel looked at that is trying to make sure if the patents cause them problems they can simply sell Intel machines and let people figure out how to add on other people's tech.
Having a larger memory set to work off means that if you know the spots are open you simply write it instead or moving or erasing what is there. Anytime you have to move over the system bus as slow as it is, you get hit with far greater latency because traveling down a wire the width of your thumb verse traveling several inches are magnitudes slower. It will still be fast perceptively much like plugging in a fiber optic cable and seeing the light on a wall next to you as the light travels down the cable... but computers are simply really fast abacuses. They can move data into and out of registers be it memory or processing units, but they are still only adding and subtracting very fast at their most basic function.
I did not see anything about the deal between Intel and AMD, but I did get to attend a private function where AMD was the main sponsor, as the newly minted Raedon group was trying to help Autodesk sell their products and sell more Fire Pro cards, as well as get people looking at where the Raedon Group wanted to be 12 to 18 months from now. They were targeting people who influence and use the workstation cards to for Engineering, Artistic Visualization, Disciples that merge those two like Architecture and Video games, and the simulation and compute crowds. It was during the compute conference so it was not surprising that they wanted to show off to people who might have to plan budgets for their companies two to four years out. So likely that is also driving the stock as they showed off where the Raedon group is aiming at. I made a joke about the larger memory addressing but until I heard that they had publicly announced it some where I was not sure how much they wanted people keeping quiet about it. Direct X is the major API people use, with open gl having some impact as there is less over head on linux machines, by default since you only load the modules you need to do the tasks you want to do.
If I say unloaded the email and notification system on windows, windows closes the gui and restarts, on linux their are a handful of systems that have to run at all times and gui is not even one of them. So if I am rendering from the command line I can use all that lost over head to render faster. If I want to see what what I am rendering I need to render on a machine with enough of the OS running to either look at the rendered output or rendering in the screen. Many people run windows because it is easier. You install the IDE and the installers create all the dependent parts. My guess is Intel looked at that is trying to make sure if the patents cause them problems they can simply sell Intel machines and let people figure out how to add on other people's tech.
I did not remember correctly. As per http://www.theregister.co.uk/2016/08/18/amd_zen_latest/ Desktop first, server second.
shad0w4life
Gawd
- Joined
- Jun 30, 2008
- Messages
- 689
MSI and gsync works if you sacrifice a small animal i can confirm.You probably have the HP laptop which for which it's known that GSync doesn't work on it. Blame HP for making a shitty laptop; blame yourself for not researching before you bought it.
Still stands that its extremely overhyped product, probably didnt help their last few driver releases have been buggy.
LOL I hated that BS when they said it. Really disappointing to me then.
Well this shows AMD business strategy being erratic as usual.
Radeon teaming up with Intel for a cut price bundle; Intel i5 6600K CPU with the MSI 480 ARMOR 8GB.
With Zen launching early next year, WTF are they doing helping push a cheap bundle with a competitor's CPU and removing a possible buyer of your own CPU out of your target audience...
http://radeon.com/en-us/radeon-intel-bundle/
And the deal: http://www.newegg.com/Product/ComboBundleDetails.aspx?ItemList=Combo.3305313
Idiots, they should be protecting the target audience to give their own product the best chance of sales.
But to me this again highlights the internal disconnect and possible conflicts of interest between VPs-senior management and the CPU-GPU divisions now split.
CEO needs to clamp down on the GPU business division; how much of this comes back to Raja, did the Intel-AMD deal involve him or was more broad because I am sure the CPU division would want something back if Intel is licensing iGPU tech.
Cheers
Radeon teaming up with Intel for a cut price bundle; Intel i5 6600K CPU with the MSI 480 ARMOR 8GB.
With Zen launching early next year, WTF are they doing helping push a cheap bundle with a competitor's CPU and removing a possible buyer of your own CPU out of your target audience...
http://radeon.com/en-us/radeon-intel-bundle/
And the deal: http://www.newegg.com/Product/ComboBundleDetails.aspx?ItemList=Combo.3305313
Idiots, they should be protecting the target audience to give their own product the best chance of sales.
But to me this again highlights the internal disconnect and possible conflicts of interest between VPs-senior management and the CPU-GPU divisions now split.
CEO needs to clamp down on the GPU business division; how much of this comes back to Raja, did the Intel-AMD deal involve him or was more broad because I am sure the CPU division would want something back if Intel is licensing iGPU tech.
Cheers
Presbytier
[H]ard|Gawd
- Joined
- Jun 21, 2016
- Messages
- 1,058
So it could be all of this is a lead up to eventually selling off the Radeon division.
If so it suggests Raja is pushing a conflict.
The CEO would not impact the CPU division sales just before they launch like this.
If there is no sell-off of the GPU division (and I am not convinced senior board management would let it go as their long term future is about integrating CPU-GPU for quite a few sectors, would come down to senior shareholders who have a stake in AMD not just the Radeon division), this is bad news from a business structural management view as these divisions need to work together, who has control of APUs (would be CPU team) and can indirectly screw back the GPU division if they so wanted to.
Any way you look at this, it is idiotic AMD business-management strategy because in the long run no-one wins.
CHeers
Last edited:
Presbytier
[H]ard|Gawd
- Joined
- Jun 21, 2016
- Messages
- 1,058
Not necessarily. let us assume for a moment the CPU team has a good handle on APU development in house and no longer needs tech from the dedicated GPU division. At that time if the companies primary concern is to build up their server business and APU sales then the Radeon division could be viewed as a drag on money and resources. AMD could be saying our best bet is to focus on one product (CPU) instead of multiple products.
Well if this is the case, its a conflict of interest, if AMD "helps" Intel get better in iGPU to an even standing with AMD's APU. This would just hurt them in the long run. As you stated it will give credence for Intel to buy RTG but also help Intel's CPU division over AMD's CPU division.
If this hurts AMD which will show up not in Kaby Lake or Cannon Lake, it will show up in a chip after that, you can bet AMD's board will look at this and heads will start rolling. I see this as a very risky move on AMD's part. This is a gamble AMD should never take,, they hold no cards. We have seen them take these risks in the past and none of them have paid off. Just the last round of the GF, is now sucking AMD dry. ATi's buyout the gamble before, just didn't work out.
Intel shit its all good for them because they will have tech to move forward to beat AMD on the CPU front and take on nV too.
And this is why I don't think this the way its going to go, because, AMD the only thing they get is money? They will be giving Intel, much more then the sum of money they would be getting from Intel. The benefit Intel gets will be something measured by a decade or two to come, which will be a hell of lot more then 100 million a quarter, even 500 million a quarter will be not enough to over come that benefit.
Contracts like this never work out, not only don't they work out, if taken to court by either party at a later date most likely will be null and voided.
If this hurts AMD which will show up not in Kaby Lake or Cannon Lake, it will show up in a chip after that, you can bet AMD's board will look at this and heads will start rolling. I see this as a very risky move on AMD's part. This is a gamble AMD should never take,, they hold no cards. We have seen them take these risks in the past and none of them have paid off. Just the last round of the GF, is now sucking AMD dry. ATi's buyout the gamble before, just didn't work out.
Intel shit its all good for them because they will have tech to move forward to beat AMD on the CPU front and take on nV too.
And this is why I don't think this the way its going to go, because, AMD the only thing they get is money? They will be giving Intel, much more then the sum of money they would be getting from Intel. The benefit Intel gets will be something measured by a decade or two to come, which will be a hell of lot more then 100 million a quarter, even 500 million a quarter will be not enough to over come that benefit.
Contracts like this never work out, not only don't they work out, if taken to court by either party at a later date most likely will be null and voided.
Last edited:
You cannot run a good CPU division without overlap and collaboration with the GPU division, what do you think Xeon Phi is?
How is the CPU team going to carry on development of Zen/new APUs-consoles of all types with the teams now split between the CPU and GPU divisions if the GPU team involved in Zen are not collaborating?
If the GPU division creates conflict, there are ways for the CPU team to respond and cause a wrinkle for them back.
What has Radeon division being viewed as a drag on money and resources got to do with it being ok that they are screwing over Zen PC gaming sales?
By that logic they would be clamped down instantly and not do a deal with Intel and the i5 as your suggesting the senior board is focused on pushing CPUs.
Anyway this is digressing that this strategy is a pain in the arse for Zen, if it is a small blip.
Cheers
Presbytier
[H]ard|Gawd
- Joined
- Jun 21, 2016
- Messages
- 1,058
I never said it made the most sense, but makes as much as the Radeon team bundling their GPUs with intel CPUs
Personally I cannot see the board or the key senior major investors allowing the GPU division to be sold off, they cannot sell their millions of shares in AMD without taking a hammering after the deal and the market makers shafting them.
And the GPU division will have staff integral to the Zen design, and a long term strategy requires good CPU-GPU tech solutions unless you have incredibly strong GPU-software-platform teams ala Nvidia (with established CUDA is within HPC tools-software) who also work with IBM;, AMD will lose out in the ARM race to Qualcomm, leaving them with x86 - APUs, and ARM development outside of the big server business.
If there ends up with a breakdown between the divisions, it is not good for any of them.
Anyway does not look a well thought out bundle deal with Intel unless designed to stick 2 fingers up at the CPU VP-division.
Cheers
Last edited:
Well that is the thing firstly Intel can't buy RTG, its impossible it will be blocked by the government. And AMD as you stated can't allow that to happen because of the integral part their GPU IP is to their APU.
Not only this even without a buyout of RTG by Intel, nV can easily sue Intel and AMD at the same time for something like this because of the ramifications it can have as having one company that holds 65%+ of the GPU marketshare joining with another that holds 15%+ of the GPU marketshare, cooperating to push out nV of the GPU market through market manipulation.
Added to this, AMD by working with Intel, giving Intel better GPU tech for their IGPU, pretty much states one of these two things
Zen is going to kill Intel, and AMD's confidence for ZEN is that high
or AMD has conceded to Intel knowing they can't compete with Intel without an influx of money and this is the gamble they are taking. Pretty much short term gains with long term ramifications just as they have done in the past. GF, Hynix, ATi, are just a few gambles that were more headaches then benefits. But this case, they could actually be bought out as a whole by another company because their bottom line might be healthy or "healthy" looking.
Not only this even without a buyout of RTG by Intel, nV can easily sue Intel and AMD at the same time for something like this because of the ramifications it can have as having one company that holds 65%+ of the GPU marketshare joining with another that holds 15%+ of the GPU marketshare, cooperating to push out nV of the GPU market through market manipulation.
Added to this, AMD by working with Intel, giving Intel better GPU tech for their IGPU, pretty much states one of these two things
Zen is going to kill Intel, and AMD's confidence for ZEN is that high
or AMD has conceded to Intel knowing they can't compete with Intel without an influx of money and this is the gamble they are taking. Pretty much short term gains with long term ramifications just as they have done in the past. GF, Hynix, ATi, are just a few gambles that were more headaches then benefits. But this case, they could actually be bought out as a whole by another company because their bottom line might be healthy or "healthy" looking.
Last edited:
I'm not sure if they want the money to compete further with intel or get money back so they can put it towards R&D, or maybe put it to actually keeping their products on shelves. They could have confidence in Zen and are using this money to give them the ability to produce more Zen CPU's for the market. I'm not sure if this money would go back into their GPU division, it seems to be taking a step back for everything else.
well competing with Intel is more money for R&D, they need to iterate Zen faster then Intel's updated schedules for their chips. At current schedules, Intel will maintain its lead (if Zen closes the gap, Intel will regain the same lead they had prior to Zen in 2 releases which is 3 years) Otherwise they won't be able to catch up. Just no way around that. On the GPU side of things, R&D was cut again this quarter, that doesn't look good when Navi is supposed to be the first "new" architecture since the 1st GCN was released........
After all this talk, when Raja stated they lost so much talent because of money, and they need to get that talent back and they cut R&D again? Sounds to me there are problems with RTG vs the rest of AMD.... Something isn't adding up.
After all this talk, when Raja stated they lost so much talent because of money, and they need to get that talent back and they cut R&D again? Sounds to me there are problems with RTG vs the rest of AMD.... Something isn't adding up.
Oh yeah also the Intel deal, which its anti competitive for AMD as well is, consoles, Intel can now build a better SOC than AMD as well :/. All of these things has to be looked into. AMD can't have Intel in their pie too. And there is no way for AMD to stop Intel from doing this because a contract that has a pharse that says Intel can't go into AMD markets because they are using AMD technologies for part of their chip, doesn't fly in court, can't stop Intel from this because a clause like that is anti competitive. If this deal that AMD and Intel are doing is really to use AMD's latest and greatest GPU tech in Intel CPU's, there will be some really big hurdles for AMD to think about in the next few years. Problems that are frankly much bigger than what they have now because it would mean Intel will be on even footing with everything AMD has and with their CPU technology and Node technology already ahead of AMD.......
Last edited: