Wednesday, May 6th 2015

AMD Readies 14 nm FinFET GPUs in 2016

At its ongoing Investor Day presentation, AMD announced that will continue to make GPUs for every segment of the market. The company is planning to leverage improvements to its Graphics CoreNext architecture for the foreseeable future, but is betting on a huge performance/Watt increase with its 2016 GPUs. The secret sauce here will be the shift to 14 nm FinFET process. It's important to note here, that AMD refrained from mentioning "14 nm," but the mention of FinFET is a reliable giveaway. AMD is expecting a 2x (100%) gain in performance/Watt over its current generation of GPUs, with the shift.

AMD's future GPUs will focus on several market inflection points, such as the arrival of CPU-efficient graphics APIs such as DirectX 12 and Vulkan, Windows 10 pulling users from Windows 7, 4K Ultra HD displays getting more affordable (perhaps even mainstream), which it believes will help it sell enough GPUs to return to profitability. The company also announced an unnamed major design win, which will take shape in this quarter, and which will hit the markets in 2016.
2015 will be far from dry for the company, as it will introduce its first GPU that will leverage HBM, offering double the memory bandwidth of GDDR5 at half its power draw. The company announced that its HBM implementation will see memory dies placed on the same package as its GPU die, so we will no longer see graphics cards with memory chips surrounding the GPU ASIC. The company will announce a few new high-end graphics cards around Computex 2015.
Add your own comment

30 Comments on AMD Readies 14 nm FinFET GPUs in 2016

#1
TheinsanegamerN
so, it sounds like 2016 will be a good year for AMD. 2015, on the other hand, is not going well at all. hopefully the r9 390x is good, because I'd love to see amd come out with both zen and the r9 400 series and re-ignite the competitive flame.
Posted on Reply
#2
Frick
Fishfaced Nincompoop
2016 will be massive. AMD stuff, and Intel 10 nm iirc...
Posted on Reply
#3
the54thvoid
Frick said:
2016 will be massive. AMD stuff, and Intel 10 nm iirc...
Zen, 10nm (iyrc), Pascal, Arctic Islands.

But..... will my system last that long......

And we need a price war.
Posted on Reply
#4
jabbadap
You sure that these aren't 16nm finfets from tsmc?
Posted on Reply
#5
Casecutter
btarunr said:
. The company will announce a few new high-end graphics cards around Computex 2015.
I guess those who circulated rumours that the release was going to be push back till HotChips in August are not correct. It really was strange such people purported a "White Paper Symposium" for engineers was going to be the event to release AMD's new line of GPU’s.


jabbadap said:
You sure that these aren't 16nm finfets from tsmc?
Well those depending on TSMC might find themselves disappointed on such a time-line by this point next year.
Posted on Reply
#6
human_error
The company also announced an unnamed major design win, which will take shape in this quarter, and which will hit the markets in 2016.
Sounds like the next Nintendo console could be the design win, as they're the only company looking at a new console in that timeframe.
Posted on Reply
#7
the54thvoid
Casecutter said:
I guess those who circulated rumours that the release was going to be push back till HotChips in August are not correct.
If you're trolling HS, he simply referenced an upcoming (later summer) conference on HBM and GPU introduction. In that same thread I'm pretty sure he didn't stick any definitive on that being a fact of release date.
To imply and falsely say (from one post, which was clarified) that he was circulating rumours is in itself, rumourtastic.

Could you refresh my memory and point to the thread? I may be wrong.

Unless of course you're not talking about HS. But you are, aren't you? because going by previous posts - he annoys you.

Don't worry, found it:

HumanSmoke said:
Well, I hope you're right. The Hot Chips Symposium could well be a post-launch whitepaper dissection, although the 16th June is a little early if you subscribe to the exactness of supposed insider knowledge. If the 16th is the drop date then we should be seeing reliable leaks (as opposed to anonymous unverified benchmark charts) fairly soon.
Posted on Reply
#8
NC37
So next year 100%, but if this year is rumored to be 60-70% then really they aren't gaining much. The major boost would be this year.
Posted on Reply
#10
Casecutter
HumanSmoke said:
If that were the case, wouldn't a more valid strategy be to point out that Fiji's launch date might be on the slide (pun intended)? The launch 390X rumours have been circulating for months - reaching a height coincidentally at the time of Titan-X's own launch, yet the Fiji and HBM official unveiling is now slated for August 25th, at the Hot Chips Symposium.
Posted on Reply
#11
the54thvoid
Casecutter said:

Okay, HS needs nobody to defend him but I don't like the pettiness this forum develops. So, for the record, the post you are quoting, preceeds my quote from HS, and I did say he clarified his point. In other words, he said what he quoted, then a few posts later, in polite conversation, posted what I quoted.

If I say "bananas are blue", then say later, "you may be right, bananas are yellow" you cant state that I circulated mistruths about the colour of bananas, when in the same conversation, I corrected that stance.

Your trolling (because that is what you are doing) is ignoring chronological positioning of statements. If statement A is corrected or clarified by statement B, quoting statement A to prove a point is itself misleading and ignorance of printed truth.

As for the actual story, it'll be a fantastic day when AMD brings it's balls to the table and proves what they can do with good direction.
Posted on Reply
#12
HumanSmoke
NC37 said:
So next year 100%, but if this year is rumored to be 60-70% then really they aren't gaining much. The major boost would be this year.
You might find that GM200/Fiji are more in line with 30-40% over the previous generation. Both Arctic Islands and Pascal should offer a considerable increase in core and overall transistor count, so 100% should be quite do-able.
the54thvoid said:
If you're trolling HS, he simply referenced an upcoming (later summer) conference on HBM and GPU introduction. In that same thread I'm pretty sure he didn't stick any definitive on that being a fact of release date.
To imply and falsely say (from one post, which was clarified) that he was circulating rumours is in itself, rumourtastic.
Could you refresh my memory and point to the thread? I may be wrong.
Unless of course you're not talking about HS. But you are, aren't you? because going by previous posts - he annoys you.
Don't worry, found it:
Right on all counts.
@Casecutter
Your quoting me does absolutely nothing to ameliorate your pathetic attempt at trolling. Firstly highlighting a question I put forward ( Fiji's launch date might be on the slide (pun intended)? ) rather than the supposed definitive statement you assume I made. A question based upon the Hot Chips Symposium where I quite correctly highlighted that Fiji and HBM's official unveiling will take place (as opposed to public launch of product). The difference you fail to understand is that product launch and white paper publication are two distinct entities - the latter generally follows the product launch -or, more rarely, is made available at the same time.
Posted on Reply
#13
Casecutter
the54thvoid said:
he clarified his point. In other words, he said what he quoted, then a few posts later, in polite conversation, posted what I quoted.
Even the most novice among should recognize what the Hot Chips Symposium is all about, or at least google and realize it’s historical underpinnings and significance to silicon chip engineering development and camaraderie is at best share/learn some of the latest developments. It's not a marketing venue.

Even after it was pointed out the response was hardly “clarified” as completely backing away from the stance.

I think members should read and decide for themselves.
http://www.techpowerup.com/forums/threads/nvidia-geforce-gtx-980-ti-silicon-marked-gm200-310.212205/#post-3276725
Posted on Reply
#14
arbiter
TheinsanegamerN said:
so, it sounds like 2016 will be a good year for AMD. 2015, on the other hand, is not going well at all. hopefully the r9 390x is good, because I'd love to see amd come out with both zen and the r9 400 series and re-ignite the competitive flame.
Frick said:
2016 will be massive. AMD stuff, and Intel 10 nm iirc...
i wouldn't count the egg's just yet as don't know what Nvidia and intel have in their tool shed. Given history Nvidia probably has something to combat it.
Posted on Reply
#15
natr0n
Wow gpu with no memory chips.

So then you would figure the cards to be much smaller...Yet concentrated heat so giant heatsinks perhaps.
Posted on Reply
#16
btarunr
Editor & Senior Moderator
human_error said:
Sounds like the next Nintendo console could be the design win, as they're the only company looking at a new console in that timeframe.
Nintendo is working on SNES-2016. A "serious gaming" console which will be 4K-ready.
Posted on Reply
#17
Frick
Fishfaced Nincompoop
arbiter said:
i wouldn't count the egg's just yet as don't know what Nvidia and intel have in their tool shed. Given history Nvidia probably has something to combat it.
We know Intel plan for 10nm in 2016 (and they, probably, recently bought them some new hardare), Nvidia is sure as europeans getting fatter planning something too. If they don't they're toast.
Posted on Reply
#19
arbiter
RCoon said:
I think members should get back on topic and talk about how awesome 14nm GPU's are (finally) instead of bickering amongst themselves.

HBM is pretty cool though. Surprised AMD were the first to bring it to the table.
AMD had to take the risk they are on their butt's atm. HBM was a risk they have to take. Nvidia will probably get on it and use it after it price comes down some and yields aren't crappy as what news says about it.
Posted on Reply
#20
alwayssts
btarunr said:
Nintendo is working on SNES-2016. A "serious gaming" console which will be 4K-ready.
Yeah...looks that way, and I agree that is probably what they're referencing, even though Nintendo said we wouldn't hear more until 2016. I wonder if that was head-fake, and/or WiiU is doing so badly they decided to move up some kind of announcement. 'Take shape this quarter' to me sounds like E3, but that seems completely infeasible...You would think they would try to get Zelda, or at least Starfox/Mario Maker out the door before talking next-gen.

I really hope whatever they come up will be on par with what we expect from the current 4k30 standard; in essence something close (in practical performance) to what 390x is. If you figure using pre-Tonga bandwidth requirements 390x would need 480Gbps (1.5x a 290x at 1018mhz), it would make sense to build a GPU or SOC around 2xHBM2 (at least 512Gbps).

Anything less seems a fool's errand if they plan to have any relevance after 'Xbox 3' and 'PS5'. IMHO it makes sense to tackle something that could feasibly scale up from current consoles and down from those in the future.

Lots of questions (even in gpu designs; reconfigured units and/or higher/lower clocks) until we know what voltages/clocks 14nm LPP will typically be able to accommodate, and how efficiently. Same goes for HBM2. Will AMD decide to go 'out of spec' again, as they appear to be doing for HBM1 (1250mhz at 1.16v vs 1ghz at 1v), or is that simply because of 28nm VDD efficiency and their current gpu design?

It seems pretty clear nvidia is shooting for 1v/256gbps x 3 for the top-end Pascal, but AMD has a tendency to not follow the spec/voltage rules if it allows a more cost-efficient and/or over-all design...and of course nvidia seems pretty content (at least with Maxwell) with supplementing (a minimum of) 1/3 of their bandwidth requirements with on-die cache...

Pretty excited to see how 14nm (and 16nm FF+ for that matter) develops. Because let's face it...this, ladies and gentlemen, is the threshold we've been waiting years and years to happen. AMD design methodologies will finally make sense wrt transistors/tdp/available bandwidth and certain performance realities will hit certain markets; cheap 1080p gpus (and igps) that don't need to make compromises, realistic 4k solutions...etc. On top of that, mobile chips (be it Kirin, 820...even Tegra) should finally be simply spectacular compared to what they need to be able to realistically accomplish.

2016 should be one hell of a year all-around.
Posted on Reply
#21
Assimilator
btarunr said:
Nintendo is working on SNES-2016. A "serious gaming" console which will be 4K-ready.
That's nice. What would be better is if they could find some original IPs for it, instead of rehashing the same old s**t for the 20th time.
Posted on Reply
#22
buggalugs
RCoon said:
.
HBM is pretty cool though. Surprised AMD were the first to bring it to the table.
Why would you be surprised AMD is first??. They were first with GDDR3, GDDR4, GDDR5 memory cards, they were also first with fully complaint DX10.1, DX11 cards, they were first with HDMI (and incremental upgrades like 1.4) and displayport ports, first to support 3 monitors on a single card, first real dual GPU cards,(post 3dfx and othe failed experiments in the 1990's) etc, pretty much all the major advancements in the last 10 years since they acquired ATI. I can only think of a couple of things when Nvidia was first like Gsync but that's just a variation of existing technology like Vsync.

There are plenty of things to criticize AMD for, but being first is not one of them.
Posted on Reply
#23
HumanSmoke
buggalugs said:
Why would you be surprised AMD is first??. They were first with GDDR3, GDDR4, GDDR5 memeory cards
Nope. That particular implementation was actually first used in Nvidia's FX 5700 Ultra.
buggalugs said:
they were also first with fully complaint DX10.1, DX11 cards, they were first with HDMI (and incremental upgrades like 1.4)
I think you'll find that the Nvidia GeForce 6600GT and 6200TC were the first cards with HDMI - vendor OEM implementations I believe.
buggalugs said:
first real dual GPU cards,(post 3dfx and othe failed experiments in the 1990's)etc, pretty much all the major advancements in the last 10 years since they acquired ATI.
ATI's Rage Fury MAXX uses/used AFR back in 1999 - six years before AMD subsumed the company.
Posted on Reply
#24
buggalugs
HumanSmoke said:
Nope. That particular implementation was actually first used in Nvidia's FX 5700 Ultra.

I think you'll find that the Nvidia GeForce 6600GT and 6200TC were the first cards with HDMI - vendor OEM implementations I believe.

ATI's Rage Fury MAXX uses/used AFR back in 1999 - six years before AMD subsumed the company.
So you're going to miss the point entirely and pick at straws? GDDR3 was designed by ATI and AMD was first with GDDR4 and GDDR5 mainstream HDMI and most of the important incremental updates like HDMI 1.4, 1.4a and soon to be HDMI 2.0 and displayport 1.2 standards on the 390X. Excluding Dual GPU cards which were experimental in the 90's.

AMD has been first with most of the major advances in GPU tech, in the last 10 years, in memory, fully complaint DX hardware and connectivity. Theres a lot of things I didn't mention too like, Driver access to overclocking, fan settings, enthusiast controls etc.

Its completely expected for AMD to be first with HBM
Posted on Reply
#25
HumanSmoke
buggalugs said:
So you're going to miss the point entirely and pick at straws?
Nope. Just correcting a couple of erroneous facts. While I agree with what you're saying as a whole, jamming in a few false claims shouldn't be hand waved away when it is just as easy to use true examples.
buggalugs said:
Its completely expected for AMD to be first with HBM
Since AMD co-developed the standard with SK Hynix, and were all but cited by Hynix as the launch customer, shouldn't that have been expected?
Posted on Reply
Add your own comment