[UPDATE] AMD Announces Polaris Architecture - 4th Generation GCN Arriving in Mid 2016

[HKZeroFive]

VIDEOCARDZ:

 

Tomorrow AMD will unveil its new architecture called Polaris. This is a major upgrade in Radeon architecture since first 28nm GPU.

Polaris is new codename for 16nm FinFET architecture that will be introduced with new graphics cards later this year. From what I’ve heard, AMD is planning to launch first Radeon 400 cards in summer. Even though the launch is still months ahead, company decided to share more details about their future portfolio to tease gamers (and probably investors).

To give you a perspective, this is how Polaris fits after 28nm architecture.

2011 — 28nm GCN 1.0 — Tahiti / Cape Verde
2013 — 28nm GCN 2.0 — Hawaii / Bonaire
2015 — 28nm GCN 3.0 — Fiji / Tonga
2016 — 16nm FinFET Polaris (GCN 4.0)

The GPU design was modified to include new logical blocks. What’s new is Command Processor, Geometry Processor, Multimedia Cores, Display Engine and upgraded L2 Cache and Memory Controller.

Last but not least, the new architecture has 4th Generation Graphics Core Next Compute Units (aka. GCN 4.0).

AMD Polaris will compete against NVIDIA’s Pascal architecture. Both are believed to utilize High-Bandwidth-Memory 2 (HBM2).

ANANDTECH:

 

As a quick preface here, while RTG demonstrated a Polaris based card in action we the press were not allowed to see the physical card or take pictures of the demonstration. Similarly, while Raja Koduri held up an unsoldered version of the GPU used in the demonstration, again we were not allowed to take any pictures. So while we can talk about what we saw, at this time it’s all we can do. I don’t think it’s unfair to say that RTG has had issues with leaks in the past, and while they wanted to confirm to the press that the GPU was real and the demonstration was real, they don’t want the public (or the competition) seeing the GPU before they’re ready to show it off. That said, I do know that RTG is at CES 2016 planning to recap Polaris as part of AMD’s overall presence, so we may yet see the GPU at CES after the embargo on this information has expired.

In any case, the GPU RTG showed off was a small GPU. And while Raja’s hand is hardly a scientifically accurate basis for size comparisons, if I had to guess I would wager it’s a bit smaller than RTG’s 28nm Cape Verde GPU or NVIDIA’s GK107 GPU, which is to say that it’s likely smaller than 120mm. This is clearly meant to be RTG’s low-end GPU, and given the evolving state of FinFET yields, I wouldn’t be surprised if this was the very first GPU design they got back from Global Foundries as its size makes it comparable to current high-end FinFET-based SoCs. In that case, it could very well also be that it will be the first GPU we see in mid-2016, though that’s just supposition on my part.

 

For their brief demonstration, RTG set up a pair of otherwise identical Core i7 systems running Star Wars Battlefront. The first system contained an early engineering sample Polaris card, while the other system had a GeForce GTX 950 installed (specific model unknown). Both systems were running at 1080p Medium settings – about right for a GTX 950 on the X-Wing map RTG used – and generally hitting the 60fps V-sync limit.

The purpose of this demonstration for RTG was threefold: to showcase that a Polaris GPU was up and running, that the small Polaris GPU in question could offer performance comparable to GTX 950, and finally to show off the energy efficiency advantage of the small Polaris GPU over current 28nm GPUs. To that end RTG also plugged each system into a power meter to measure the total system power at the wall. In the live press demonstration we saw the Polaris system average 88.1W while the GTX 950 system averaged 150W. Meanwhile in RTG’s own official lab tests (and used in the slide above) they measured 86W and 140W respectively.

 

Thankfully for Polaris, RTG is revising their naming policies in order to present a clearer technical message about the architecture. Beginning with Polaris, RTG will be using Polaris as something of an umbrella architecture name – what RTG calls a macro-architecture – meant to encompass several aspects of the GPU. The end result is that the Polaris architecture name isn’t all that far removed from what would traditionally be the development family codenames (e.g. Evergreen, Southern Islands, etc), but with any luck we should be seeing more consistent messaging from RTG and we can avoid needing to create unofficial version numbers to try to communicate the architecture.

To that end the Polaris architecture will encompass a few things: the fourth generation Graphics Core Next core architecture, RTG's updated display and video encode/blocks, and the next generation of RTG's memory and power controllers. Each of these blocks is considered a seperate IP by RTG, and as a result they can and do mix and match various versions of these blocks across different GPUs, such as the GCN 1.2 based Fiji containing an HEVC decoder but not the GCN 1.2 based Tonga. This, consequently, is part of the reason why AMD has always been slightly uneasy about our unofficial naming. What remains to be seen then is how (if all all) RTG goes about communicating any changes should they update any of these blocks on future parts, and whether say a smaller update like a new video decoder would warrant a new architecture name.

 

For GCN 4, RTG will be making a number of changes in order to improve the efficiency of the architecture from both a throughput perspective and a power perspective. Of the few details revealed to us, one of the things we have been reassured of is that GCN 4 is still very much Graphics Core Next. I expect that means retaining the current 16-wide vector SIMD as the base element of the architecture, in which case it should be fair to say that while GCN 4 is the biggest change to GCN since its introduction, RTG is further building on top of GCN rather than throwing large parts out.

As a result a lot of the disclosed GCN 4 key features point to improving throughput of GCN. Improved shader efficiency is somewhat self-explanatory in that regard. At the same time RTG is also disclosing that we will be seeing some kind of hardware scheduler in GCN 4 along with instruction pre-fetch capabilities, which again should help them improve throughput in ways to be determined. Meanwhile in an improvement for the GPU front-end, GCN 4 will be adding a primitive discard accelerator – and again we don’t have any further details than this – but it should help the architecture clamp down on getting rid of unseen geometry. Finally, GCN 4 will also include a newer generation of RTG’s memory compression technology, this coming just one generation after it was most recently (and significantly) improved for the third generation GCN architecture (GCN 1.2).

 

Finally, along with today’s announcement RTG is also disclosing a bit of what we can expect for Polaris’s multimedia controllers. On the decode side RTG has confirmed that Polaris will include an even more capable video decoder than Fiji; along with existing format support for 8-bit HEVC Main profile content, Polaris’s decoder will support 10-bit HEVC Main10 profile content. This goes hand-in-hand with the earlier visual technologies announcement, as 10-bit encoding will be necessary to prevent banding and other artifacts with the wider color spaces being used for HDR. Meanwhile RTG’s UVD video encode block has also been updated, and HEVC encoding at up to 4K@60fps will now be supported, marking a major jump over the previous generation encoder block only supporting H.264. There is no word on whether that will include Main10 support as well, however.

Meanwhile RTG has also disclosed that the first Polaris parts are GDDR5 based. Going hand-in-hand with what I mentioned earlier about RTG’s Polaris demonstration, it seems likely that this means we’ll see the lower-end Polaris parts first, with high-end parts to follow.

 

As for the specific laptop and desktop markets, RTG tells us that desktop and mobile Polaris parts will launch close together. I wouldn’t be surprised if mobile is RTG’s primary focus since it’s already the majority of PC sales, but even so desktop users shouldn’t be far behind.

The group has confirmed that they will be utilizing both traditional partner TSMC’s 16nm process and AMD fab spin-off (and Samsung licensee) GlobalFoundries’ 14nm process, making this the first time that AMD’s graphics group has used more than a single fab. To be clear here there’s no expectation that RTG will be dual-sourcing – having both fabs produce the same GPU – but rather the implication is that designs will be split between the two fabs. To that end we know that the small Polaris GPU that RTG previewed will be produced by GlobalFoundries on their 14nm process, meanwhile it remains to be seen how the rest of RTG’s Polaris GPUs will be split between the fabs.

 

VIDEO:

 

 

Huh, I didn't know that GCN 2.0 and 3.0 versions existed. From the slides, it's hard to see if it's either 16nm or 14nm architecture as there's no clarification on that. The article claims that this announcement will be made in a press release tomorrow so I will update this as soon as possible whenever possible. Exciting news nonetheless as it provides a rough idea of what AMD's "Arctic Islands" will shape up to be, to rival NVIDIA's Pascal.

EDIT: To clear up the confusion about GCN 2.0 and 3.0, AMD has always referred to them as just GCN (no version).They have never used the naming scheme 1.x to describe them. The press adopted the 1.x naming pattern because there was little evolution of the architecture so that just stuck to the public. I edited the title to 4th Generation GCN to clear this up.

TL;DR Version

 

-New GPU Core Architecture (duh)

-Significant emphasis on power usage reduction from the previous 28nm, as seen in the live demonstration of Battlefront between the Polaris system and GTX 950 system (possible bias?)

-16/14nm FinFET confirmed

-Support for HDMI 2.0a, DisplayPort 1.3 and 4K h.265

-Polaris uses both GDDR5(X?) and HBM(2?) - higher cards using HBM

-Low/mid range cards releasing first, high end cards releasing later

 

UPDATE: More slides. The slides are apparently under embargo until sometime today so expect press releases soon.

UPDATE 2: It's official.

Sauces:

http://videocardz.com/58031/amd-announces-polaris-architecture-gcn-4-0

http://www.anandtech.com/show/9886/amd-reveals-polaris-gpu-architecture

[ivan134]

What happened to 2.0 and 3.0?

 

EDIT: Oh, wtf? GCN 1.1 and 1.2 are 2.0 and 3.0 now? Since when?

[patrickjp93]

4.0? Where the f*ck are 2.0 and 3.0?

 

*Sees source*

 

Oh, Videocardz? If even wccf isn't reporting on it, it's bogus.

[Chasem121]

Looks like it will be an exciting year for graphics cards

[AnnoyedShelf]

What happened to 2.0 and 3.0?

 

EDIT: Oh, wtf? GCN 1.1 and 1.2 are 2.0 and 3.0 now? Since when?

 

4.0? Where the f*ck are 2.0 and 3.0?

 

*Sees source*

 

Oh, Videocardz? If even wccf isn't reporting on it, it's bogus.

R9 200 was 2.0

R9 300 was 3.0

[-rascal-]

This news source is a fail...

[HKZeroFive]

What happened to 2.0 and 3.0?

 

Officially Tahiti is GCN 1.0, Bonaire is GCN 2.0 and Fiji is GCN 3.0, making Polaris GCN 4.0.

 

I read that the press uses v1.0, v1.1 and v1.2 because the differences were so small. I guess AMD didn't want to name it that way.

[The_Stig]

Yay, an improvement over what we had 5 years ago. Took them long enough

[BuckGup]

Yeah!

[TetraSky]

Well it's more like GNC 4th Gen and not 4.0...

[SurvivorNVL]

This is all well and good, but I would preferably like to see if we can now drive 5k on a single GPU.  I would also like some 5K HDR monitors and a sandwhich.  

[MEC-777]

Not really much info to go on and source is sketchy. I'm more interested to know which tier cards will use HBM2 and which will adopt GDDR5X Vram.

[Spongy141]

When is it expected to arrive?

[Mew]

So I can assume this new architecture won't be good because they've been alternating between good and bad every generation-

tahiti- good

hawaii- bad

fiji- good

polaris- bad?

[Samfisher]

R9 200 was 2.0

R9 300 was 3.0

You mean they renamed the cards and they got upgraded magically to a whole new version? Kek.

[SurvivorNVL]

Not really much info to go on and source is sketchy. I'm more interested to know which tier cards will use HBM2 and which will adopt GDDR5X Vram.

I kind of would rather them all use GDDR5X / GDDR6 as it'll be plenty of bandwidth and lower latency than HBM, unless there're some latency tweaks to HBM in HBM2.

[patrickjp93]

I kind of would rather them all use GDDR5X / GDDR6 as it'll be plenty of bandwidth and lower latency than HBM, unless there're some latency tweaks to HBM in HBM2.

 

Eh, I kinda doubt HBM 2 is really going to lower latency all that much. There has to be interconnect overhead involved with chip stacking. Doubling the layer count AND lowering latency? I don't buy it until I see proof.

[SurvivorNVL]

Eh, I kinda doubt HBM 2 is really going to lower latency all that much. There has to be interconnect overhead involved with chip stacking. Doubling the layer count AND lowering latency? I don't buy it until I see proof.

Isn't nVidia onboard to make use of HMC after Pascal?  Isn't Volta utilizing it or is that only for data-center applications?

[AnnoyedShelf]

You mean they renamed the cards and they got upgraded magically to a whole new version? Kek.

Not entirely.  There were a few cards in each series that got 2.0 or 3.0 respectively.

[patrickjp93]

Isn't nVidia onboard to make use of HMC after Pascal?  Isn't Volta utilizing it or is that only for data-center applications?

 

Were*

 

It will depend on how HMC evolves after Pascal. I kinda doubt it will catch up with HBM 2 in bandwidth per stack, but we'll see.

[Kobz360]

So I can assume this new architecture won't be good because they've been alternating between good and bad every generation-

tahiti- good

hawaii- bad

tonga- good

polaris- bad?

Hawaii was bad? The 290/390/X have definitely stood the test of time and remain to this day, to be highly competitive cards. 

[patrickjp93]

Hawaii was bad? The 290/390/X have definitely stood the test of time and remain to this day, to be highly competitive cards. 

 

They compete with dragons on who spits more fire.

 

Hurr Durr hot cards Hurr Durr.

[Kobz360]

They compete with dragons on who spits more fire.

Hurr Durr hot cards Hurr Durr.

Big P mate, that's savage

[SurvivorNVL]

Polaris = GCN

Arctic Islands = GPUs based on Polaris

Greenland = High End Cards in in AI which is in Polaris

I still can't believe it is not butter.

[patrickjp93]

Big P mate, that's savage

 

Not as savage as that 1-post troll who put up that Global Foundries lost its New York 14nm fab.

 

Gotta admit, that was pretty good putting up a picture of a fire at night and faking the link text back to the NYT. I kinda wish the mods hadn't erased the thread so fast. As much as I lost my shit and cussed, that was pretty good. I'm pretty sure he'll show up on AMD reddit at some point. I just...I'm seven beers and two shots of Absolute into the night and I saw that and all Hell broke loose.

 

@MoonSpot, @Clear17Mud, and @fire219 all saw that. I didn't get to see 2 of the comments though. That was just so bad. Waterworks and everything for the first 3 comments and I flipped out. My Dad thought I was nuts for a split second. Just, wow...