![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Fury won't beat 980Ti, much less Titan
- Thread starter Stoly
- Start date
- Status
Nah, I don't think they are done in CPUs. Their desktop CPUs aren't competitive but the newest APUs are promising, and integrated all-in-one CPUs are where the market is heading. They just need a better fab. Intel is killing everyone with their process advantage.
I don't really have any expectations, honestly. I think that they are taking a smarter direction with the design of the CPU based on what has leaked, and with the CPU intended to be on 14nm process I think that it will be a lot more competitive on that alone. It's hard to compete with 14nm Intel chips on a 28nm process.If you read into that Zen should be "close" I agree if you are thinking that Zen should dominate the x86 market then you are wrong.
If it was easily possible to have such a big leap forward from a company which has the deck stacked against them then Intel would be really bad at making cpu and there not.
TaintedSquirrel
[H]F Junkie
- Joined
- Aug 5, 2013
- Messages
- 12,691
Does AMD even have the resources to compete with Intel anymore?
It's like trying to win the Daytona 500 on a unicycle. I can hold press conferences and talk about how my unicycle is "special" and comes equipped with some new kind of super wheels that will rocket me to the #1 spot. Still gonna come in last place.
And that's starting to happen in the GPU market.
Intel spends a huge amount of money on R&D but a lot of that goes towards their fab technology, it's not all spent on CPU design.
Do I think AMD is going to leapfrog Intel technologically? No.
Do I think that AMD could still put out a competitive x86 design? Yes.
I agree that integrated CPU and GPU are important. It seems AMD invested a lot in HSA and really got nothing out of it. I just think you have to get out of the "commodity" markets. Intel owns those. Custom SoC, etc. could be profitable. Especially if the CPU part is used to leverage the GPU strengths (sort of what nvidia is doing).
Targeting notebooks, desktops, laptops, even enterprise stuff -- all a waste of time. Intel owns all of that up and down, and some of those markets have low margins as it is. And even AMD is forecasting negative growth in many of those markets.
Targeting notebooks, desktops, laptops, even enterprise stuff -- all a waste of time. Intel owns all of that up and down, and some of those markets have low margins as it is. And even AMD is forecasting negative growth in many of those markets.
durquavian
Gawd
- Joined
- Dec 13, 2014
- Messages
- 757
HSA starts with W10. Not sure if it is actually out but have seen a few claim it was working with W10 by watching the usage while equipped with dGPU.
HPC is kind of up in the air at this point. Nvidia used to talk about it every chance they got and now they're kind of mum. I think people see FPGAs as taking over in that area.
The best benefits should be in the APU area. Running on the dGPU could be done prior to Windows 10 and is sort of non-revolutionary. It's just glorified OpenCL / CUDA, etc. The whole thing is a mess, really.
durquavian
Gawd
- Joined
- Dec 13, 2014
- Messages
- 757
Didn't mean on dGPUS but on iGPU with dGPU equipped.
defaultluser
[H]F Junkie
- Joined
- Jan 14, 2006
- Messages
- 14,398
No, Custom SoC design is only profitable if you have a unique product + business plan, which AMD definitely does not. If you do it on-contract, you're stuck fighting with Chinese chip designers that can cut every corner known to man, and giants like Intel\/Qualcomm/Samsung who can sell their low-end products at a loss. See how Nvidia got kicked out of phones and mainstream tablets by Qualcomm, Intel, Rockchip and Mediatek? That happened because Nvidia are INEFFICIENT compared to the Chinese, and can you imagine how much more inefficient AMD is (consider which company is shrinking)? Nvidia is at least trying to re-purpose their core design in other markets (gaming, car entertainment), which is something AMD would never have the balls to do. You need to have that kind of flexibility to compete in the open market.
High-end game consoles aren't redesigned often enough for AMD to make that their primary business, and nearly everything else will just use mainstream ARM or x86 chips, and paired with discrete GPUs when necessary. There's hardly any need for chip design houses anymore that don't build and sell their own products, since the market is so volatile and cheap.
Since the PC market is now shrinking, as has the discrete GPU market for half a decade now, I see no out for AMD. They can't out-market and out-research Nvidia, and they can't out-anything against Intel.
Last edited:
Pieter3dnow
Supreme [H]ardness
- Joined
- Jul 29, 2009
- Messages
- 6,784
The problem was the same as what they had with the Phenom II they said that they could not get any more out of it in any way shape or form.
To not have to rely on compilers or MS sad attempts at processor scheduling will certainly be a plus for Zen..
Pieter3dnow
Supreme [H]ardness
- Joined
- Jul 29, 2009
- Messages
- 6,784
To be honest Nvidia was never into phones and tablets in the first place.
Consoles are just a good way to generate money for AMD. They never said it would be their primary business and they prolly will never do that. They have some obligations towards global foundries and that might also help out a little ...
But you forget about Qualcomm is that they use AMD technology.
for whatever reason Nvidia is getting a bit restless when trying to litigate:
The ongoing twists to the Nvidia patent trolling ‘Kepler license’ scheme/
http://semiaccurate.com/2015/05/11/nvidias-q12016-analyst-call-three-unusual-themes/
And that is what Nvidia is up to exactly.
At this point in time Intel is trying to get a foothold into the mobile market, which is about 3 years to late.
If you want to check you will find out what is powering those Apple mobile devices the guy/team which did the cpu for Apple works for AMD now and is working on the ARM version.And AMD seems to be the GPU Apple uses these days.
Unique product and business plan sounds very logical to you but it is so far out there that I'm wondering if you know anything about the processor market.
OFaceSIG
2[H]4U
- Joined
- Aug 31, 2009
- Messages
- 4,043
This. AMD has done this time and time again and it's sold cards. If it's 90% the performance of 980 Ti for 60 or 70% the price it's a win. We all know that's how the runner up in this sector works.
TaintedSquirrel
[H]F Junkie
- Joined
- Aug 5, 2013
- Messages
- 12,691
No way they can do this with costs from HBM and that big ass die.
Unless they get a huge discount from Hynix. They did co-develop.
If they can pull another 290 ($399) vs 780 ($650) with Fiji, I will literally lose my shit.
durquavian
Gawd
- Joined
- Dec 13, 2014
- Messages
- 757
Wait... Found it. what did you have for dinner...
defaultluser
[H]F Junkie
- Joined
- Jan 14, 2006
- Messages
- 14,398
I guess your memory is rusty. Luckily the internet never forgets!
Nvidia Tegra 2 is sole launch platform for Android Ginerbread (i.e. TABLETS), and also launched a handful of phones:
http://www.anandtech.com/show/4117/motorola-xoom-nvidia-tegra-2-the-honeycomb-platform
And Tegra 3 released a year later, and was used by a ton of Android tablets, including the upgraded Asus Transformer, the ORIGINAL Nexus 7, Microsoft Surface, Android tablets from Sony, Toshiba, Acer, HP. Here is the fist review of a Tegra 3 tablet:
http://www.anandtech.com/show/5163/asus-eee-pad-transformer-prime-nvidia-tegra-3-review
Nvidia having not tablet aspirations? They practically built the Android tablet market themselves! And then they fucked-up Tegra 4 and 4l, and left the tablet market to Qualcomm, Intel and China.
Nope. Acquired with the ATI purchase, and tossed aside after years of neglect by AMD management. They wanted no part in the mobile revolution, and sold it entirely to Qualcomm, with all models beyond the Adreno 200 internally developed from the base technology. They get no license fees for this:
http://www.digitaltrends.com/mobile/qualcomm-buys-amdatis-mobile-graphics/
The only thing Apple cares about besides battery life is "premium" unique features they can sell to the idiots, like retina screens and PCIe storage. GPUs are so passe, they just take whatever they can get the best deal on that just happens to be better performing than their last chip, and fits their power budget. You and I both know they didn't exactly go spend-happy on the GTX 650/750M models in the previous two years.
You can be sure that AMD doesn't get even a fraction of the $2500 asking price for the 15" Macbook Pro. This is why they are doomed: they can't charge a premium on components when nobody needs top-performance anymore, and they don't have a plan to package those old things in a new sexy package like Apple.
And next year Apple will announce with much aplomb that they've upgraded to the CUTTING-EDGE Nvidia 950M
Last edited:
TaintedSquirrel
[H]F Junkie
- Joined
- Aug 5, 2013
- Messages
- 12,691
Five Guys.
LurkerLito
2[H]4U
- Joined
- Dec 5, 2007
- Messages
- 2,706
Personally I think this is a trial run for Apple, and if they like the quality AMD offers, at a < $3 price on the stock price, I wouldn't be surprised if Apple just buys out AMD and makes it exclusive to Apple products. It would solve their problems with using Intel cpus and any other provider for the graphics.And next year Apple will announce with much aplomb that they've upgraded to the CUTTING-EDGE Nvidia 950M
Samsung is fabbing chips for nvidia as well, so I don't think there's much smoke there. In fact the deal with AMD / GF was before nvidia's lawsuit anyway, IIRC.
Why would Apple want to take on the albatross that is AMD? It could literally be the downfall of their company.
primetime
Supreme [H]ardness
- Joined
- Aug 17, 2005
- Messages
- 7,338
LurkerLito
2[H]4U
- Joined
- Dec 5, 2007
- Messages
- 2,706
It really depends on what Apple wants to do. Apple has expressed interest in getting away from Intel, they can't buy a Qualcomm or Arm, but it would be possible to get a lot of IP from buying AMD. Then they could tweak the company to put more emphasis on power savings and all the other things Apple needs. I would be surprised if they did it because AMD doesn't have a Arm like equivalent, but it would be a inexpensive solution for buying a company geared towards making customized processors for their products if they can tweak it inexpensively.
dammit, now i want a five-guys bacon-cheese all-the-way.
I don't know where AMD goes from here. They obviously don't have the R&D budget to compete with Nvidia anymore. I was hoping that by them taking a risk with HBM they'd be able to one up Nvidia, but the card is coming way too late. Nvidia will have 16nmFF+/HBMv2/Pascal cards out by March most likely. So anything AMD stands to gain here will just be lost then. And honestly if the performance rumors are true that they're desperately trying to increase clocks last minute because it's not even as fast as a stock 980 Ti - what does AMD stand to gain here anyway? And the thing won't have any overclocking headroom left because AMD will have already used it all up to keep up with stock Nvidia flagships. Stock liquid cooling is just a temporary solution to a potentially long-term problem and ultimately a warning sign.
Most people will probably just buy the 980Ti -- which is already out. The Fiji card is rumored to have production issues regardless due to the HBM so supply is going to be short for while.
It's kinda sad -- because AMD is in essence paving the way for it's own destruction. They basically are doing all the leg work for HBM and Nvidia is going to be the one reaping the benefits with V2.
Dark days for team Red.
Most people will probably just buy the 980Ti -- which is already out. The Fiji card is rumored to have production issues regardless due to the HBM so supply is going to be short for while.
It's kinda sad -- because AMD is in essence paving the way for it's own destruction. They basically are doing all the leg work for HBM and Nvidia is going to be the one reaping the benefits with V2.
Dark days for team Red.
Last edited:
TaintedSquirrel
[H]F Junkie
- Joined
- Aug 5, 2013
- Messages
- 12,691
Maxwell's release to now is the same amount of time as now to March.
How is AMD late now but March 2016 is soon???
Time is an Nvidia fanboy apparently...
How is AMD late now but March 2016 is soon???
Time is an Nvidia fanboy apparently...
It's not obvious that the x86 license would even transfer, and Apple already has a handle on in-house ARM designs. But more importantly, top technology isn't really a priority for Apple. They don't care about having the fastest this or most feature-full that. They sell fashion and image. They sell the idea that you're better than everyone else because you bought their product. Kind of like Mercedes, BMW, etc. Apple takes the lowest-level, basement-bargain, cheapest crap out there and slaps their logo sticker on it and that's how they have such amazing margins. Absorbing AMD into the company could hurt that image. AMD's public image is the "budget brand" for poor people. It would be like BMW buying Pontiac.
Titan X is about 3-5% faster than 980 Ti. So if AMD aimed to beat Titan X performance (even if just at 4K) then a card that performs marginally worse than Titan X isn't going to throw them for a loop. Titan X was released in March, so the only assumption then is AMD just sat on their asses for 3 months if they're still adjusting clocks to beat a slower 980 Ti, or rumors are just rumors.
As far as 14/16nm FinFET GPUs go, even Intel with the world's most advanced fabs had significant trouble with the 14nm node. The first Broadwell Core M parts were <100mm^2 dies (82mm^2 to be exact), and only just a few days ago did the >100mm^2 5775C and 5675C launch. And as you increase your die size, yield starts going down, and not linearly either. Until Intel manages volume production of >300mm^2 dies (likely for 12+ core Xeon parts), we can forget about 14/16nm GPUs, because last I checked TSMC is no Intel. Then throw in the fact HBM2 is another wild card, I highly doubt we'll see an HBM2 equipped >300mm^2 Pascal GPU before Q3 2016.
Pieter3dnow
Supreme [H]ardness
- Joined
- Jul 29, 2009
- Messages
- 6,784
You can be sure that AMD doesn't get even a fraction of the $2500 asking price for the 15" Macbook Pro. This is why they are doomed: they can't charge a premium on components when nobody needs top-performance anymore, and they don't have a plan to package those old things in a new sexy package like Apple.
And next year Apple will announce with much aplomb that they've upgraded to the CUTTING-EDGE Nvidia 950M
Nvidia never had a foothold I'm sure they brought products to the tablet market but they are nothing but expensive hand warmers ....
You make it sound like margins are everything. If you can't sell things what impact do better margins have, zero.
Surely they could buy qualcomm AND arm they could even throw mediatek to the mix and keep the change.
But apple has already invested a lot in its own ARM technology and I don't think qualcomm would add a lot of value.
LurkerLito
2[H]4U
- Joined
- Dec 5, 2007
- Messages
- 2,706
No Qualcomm and Arm would be expensive to buy. Qualcomm has a $111B market cap Arm is like $44B and AMD is less than $2B. IIRC Apple has like $150-$160B in cash so AMD would be least expensive to buy for IP holdings.
YeuEmMaiMai
Extremely [H]
- Joined
- Jun 11, 2004
- Messages
- 34,173
do you all really want apple to buy AMD? if they did, you would pay even more for an accelerator than you would if Nvidia was all alone in the market,,,,
The most anticipated new product from AMD is Fury, and following what has been previously mentioned here, the Fury lineup will consist of three models. These models are Fury Nano, Fury XT, and Fury Pro.
•The Fury series will first come in reference models and both water-cooled and air-cooled products will be released. Non-reference models will come to market as early as mid-August or as late as early September. For the next month or more, only reference models will be released.
•As expected, the supply of reference models will be limited.
•We believe that the AMD Fury series is sufficiently competitive with the TITAN X and the GTX 980 Ti.
•AMD Fury will have a GPU and memory on top of a small interposer, so the concentrated heat from the interposer region looks unsatisfactory concerning the TDP, but in the reference models this problem is limited.
•The heat from the cooler is of a high level (like the TITAN X), but the noise level is of no concern.
•The Fury series will first come in reference models and both water-cooled and air-cooled products will be released. Non-reference models will come to market as early as mid-August or as late as early September. For the next month or more, only reference models will be released.
•As expected, the supply of reference models will be limited.
•We believe that the AMD Fury series is sufficiently competitive with the TITAN X and the GTX 980 Ti.
•AMD Fury will have a GPU and memory on top of a small interposer, so the concentrated heat from the interposer region looks unsatisfactory concerning the TDP, but in the reference models this problem is limited.
•The heat from the cooler is of a high level (like the TITAN X), but the noise level is of no concern.
TaintedSquirrel
[H]F Junkie
- Joined
- Aug 5, 2013
- Messages
- 12,691
I have a hard time believing AMD is going to release both air and water ref cards.
Mid-august for custom cards... Better hope that's wrong.
Mid-august for custom cards... Better hope that's wrong.
LurkerLito
2[H]4U
- Joined
- Dec 5, 2007
- Messages
- 2,706
No one really wants Apple to buy AMD, it's just an idea that sounds interesting if you think about it from the point of view of Apple. If they really wanted to make their own chips they could easily buy AMD given it's marketcap just for it's IP. They could buy Nvidia also without a problem but it would cost them 10x the price, but AMD would be more in line to the type of chips they would want to build, CPU and GPU, while Nvidia would only net them GPU, though I guess they also have Tegra but I doubt that is of much interest for Apple.
DX12, win10 and such also are key points as AMD has bet on the future of gaming.
Warstopper
Weaksauce
- Joined
- Nov 14, 2013
- Messages
- 67
Apple would be a terrible place for AMD to end up. I had always hoped IBM would purchase AMD but IBM is worthless now. Samsung would be a much better partner for AMD and I think they could package AMD and Global Foundries in order to match Intel's capacity.
That said, I don't think either company would benefit the industry as a whole.
I think some interesting choices would be Oracle, Lenovo or (gag) Cisco as they would drive performance and efficiency in the server chip which would transition over to the consumer side. I also think Oracle would push software through that benefited the APU and ultimately showed them in a much better light than we currently see (ie, HSA). But then they would be owned by Oracle
Either way, I look forward to the coming Fury product and I really hope they do something special. I always purchase AMD as I feel we need them in order to keep intel from charging nearly 700 for a Pentium 60 again.
That said, I don't think either company would benefit the industry as a whole.
I think some interesting choices would be Oracle, Lenovo or (gag) Cisco as they would drive performance and efficiency in the server chip which would transition over to the consumer side. I also think Oracle would push software through that benefited the APU and ultimately showed them in a much better light than we currently see (ie, HSA). But then they would be owned by Oracle
Either way, I look forward to the coming Fury product and I really hope they do something special. I always purchase AMD as I feel we need them in order to keep intel from charging nearly 700 for a Pentium 60 again.
No, this is wrong. Apple uses quality hardware and delivers optimal customer service--as survey after survey after survey demonstrates. The most expense to a company is customer service so it's clear that your understanding of the market is a bit off, to say the least.
The reason Apple has such high margins is because they have successfully built a premium brand, as you mention BMW and Mercedes it's obvious you haven't ever been in either vehicle because they charge premium prices for premium quality. You don't even have to drive in one to notice the qualitative differences between a premium European brand and a Honda--simply open and close a door or trunk and the differences are stark (assuming you know what to look or listen for...although if your entire life has been dictated by budget constraints you might not know what to look for similarly for someone who's only ever tasted gut rot alcohol asked to differentiate between premium brands of spirits). Marketing only gets a company so far, they have to deliver on the promises or customers will balk and walk.
Apple's high margins aren't based on using the cheapest components, but rather based on using the cheapest manufacturing techniques. They have lean product lines, which cut costs down substantially. They have guaranteed market share, which allows them to dictate favorable terms. They also utilize pull to market, rather than push. In basic terms, this means they generate market demand while their customers wait for the product to arrive rather than producing stock that has to sit on shelves and waste valuable (expensive) store space.
When business students learn about Apple strategies they don't read about taking a 5cent part and hyping it to sell for $100 dollars. That'd be a stupid strategy most of all--good enough for a one-off company but hardly a model of sustainable profits like Apple shows from year to year to year to year, etc. Rather they learn how to drive efficiencies from a production standpoint, since that's where companies can compress costs most. That's been true since at least Henry Ford, who didn't invent the automobile but was able to revolutionize production techniques.
Many will write off my comments as some kind of fandom but I have my doubts about who's running multi-billion dollar companies here. I don't run any either, but I at least read the books written by the guys who did it and their theories about why they did it so I have some kind of understanding of what people mean when they relate Jobs to Ford and talk about revolutionizing the market space instead of relegating it to puerile comments about rounded corners and snake oil claims.
- Status