AMD 14nm FinFET Polaris GPU Size Leaks Out – 232mm² Large Die

[Mr_Troll]

AMD’s Polaris die with 232mm² surface area spotted
 

Quote

Details of AMD’s upcoming 14nm GPU have finally started leaking out. Some good detective work by the user AnarchX over at the Beyond3D forums has revealed what appears to be the LinkedIn profile of a senior engineer at AMD. Interestingly, the engineer lists multiple projects, one of which is the Polaris die. The size of the chip will be 232mm² and (assuming the information is accurate) will constitute one tier of the Polaris architecture

 

Quote

 

Lets start with the customary context. AMD revealed its 14nm based Polaris lineup some time ago and the fact that it will be based on a FinFET process. Initially AMD did not clearly mention whether the process would be 14nm or 16nm, something which was later clarified as being only 14nm (for Polaris architecture). We also learned that AMD is focusing on bringing the ‘minimum spec VR’ level of performance (think R9 290 or a GTX 970) to more affordable levels and within reach of more gamers. All this ties in with the Polaris strategy – which is broken down into Polaris 10 and Polaris 11.

We know that the chip in question is a Polaris die because 1) It is based on the Samsung/GloFo 14nmFinFET LPP node, which is exactly the right node for the Polaris GPU and 2) it says dGPU and AMD isn’t working on any other 14nm GPU right now. Before we go any further, given below is the relevant portion from the engineer’s LinkedIn profile:

 

 

Quote

dGPU project F: (Global Foundry/SAMSUNG 14LPP, 430 blocks, 232mm2)
As full chip floorplan designer in Top leader team, take responsibility of manual feedthrough insertion, and lead pin assign & repeater insertion, etc, project is ongoing. Be familiar with latest leading technology.

Quote

According to the information we have about the 14nm LPP process, and based on transistor density increase, a 232mm² GPU would be roughly equivalent to a  464mm² 28nm processor – at the same TDP levels. Since we already know that AMD is going to be focusing not just on performance but power efficiency as well – this number could be be much higher, in fact we will discuss the number AMD is using below. We can however safely say that this die is more than capable of meeting the minimum VR spec that AMD promises.

 

A 232mm² die multiplied by 2.5 times gives us the performance range of a 28nm 580mm² chip (at the same TDP level) which would be pretty damn impressive. We do not know what kind of power efficiency mix AMD is playing with here so it would be pointless to speculate further about the exact performance levels at this point. It would however, be prudent to keep in mind the difference between a 28nm and 14nm die.

 

At any rate, here are the two possibilities with this leak: 1) The information revealed is simply of a chip design that never went on to see the light of the day (the engineer just worked in the building of the floorplan design – which is a long way from taping out) or 2) we are looking at the real deal and something which will target new levels of performance/watt and affordability to bring VR-Ready performance levels to more gamers. If the latter is true than we could see an even powerful GPU from AMD further down the road.

 

 

Can you leak some benchmarks amd pls?

 

Source:https://www.linkedin.com/in/jurgen-hao-8483a768

http://www.forum-3dcenter.org/vbulletin/showthread.php?p=10939935#post10939935

http://wccftech.com/amd-polaris-gpu-die-size-232-mm-2/

 

 

[Dabombinable]

Can we hope for a GPU with the die size of, for example, that of something such as the Geforce 6200.

[zMeul]

if something designed reaches engineering state, does not mean it will end up as a end product

[stconquest]

Is Polaris introducing CUDA support on AMD?

[PlayStation 2]

12 minutes ago, zMeul said:

if something designed reaches engineering state, does not mean it will end up as a end product

Chances are, we're at the point where late test samples are already developed. Polaris as a whole was probably taped out in July.

[zMeul]

4 minutes ago, Dan Castellaneta said:

Chances are, we're at the point where late test samples are already developed. Polaris as a whole was probably finished in July.

as in taped out?! then, yes

[PlayStation 2]

Just now, zMeul said:

as in taped out?! then, yes

Yeah, should have been a bit more specific there.

[Notional]

11 minutes ago, stconquest said:

Is Polaris introducing CUDA support on AMD?

No they probably never will either. Especially not when they created a way to easily convert CUDA code to c for compute.

 

That being said, this looks amazing. This chip would fit perfectly in a laptop or tiny card for tiny builds.

[byalexandr]

Just imagine.... Polaris cards more powerful than the R9 Nano but just as short and half height....

[Coaxialgamer]

I really hope this is the small die implementation of polaris.Let's hope polaris 11 has a big die. Otherwise , i'm waiting for the 590x . Ugh.

[DocSwag]

I'll be upping my gpu at the end of 2017... If this isn't a top end gpu looks like I'm looking at the 590x/590 like coaxialgamer

[Jury]

Can't wait for Polaris. This 14nm will really help me on my ridiculous power bill.

[Khvarrioiren]

2 hours ago, Coaxialgamer said:

I really hope this is the small die implementation of polaris.Let's hope polaris 11 has a big die. Otherwise , i'm waiting for the 590x . Ugh.

Keep in mind these are the first 14nm cards out there. There are bound to be bigger die size GPU's coming along the way.

[Coaxialgamer]

Just now, Khvarrioiren said:

Keep in mind these are the first 14nm cards out there. There are bound to be bigger die size GPU's coming along the way.

i know , but they ARE pretty small.

[SuperCookie78]

2 hours ago, Notional said:

No they probably never will either. Especially not when they created a way to easily convert CUDA code to c for compute.

 

That being said, this looks amazing. This chip would fit perfectly in a laptop or tiny card for tiny builds.

I thought AMD finnaly  decided to get a CUDA license. 

[azurezeed]

that looks promising, let's hope AMD delivers

[Notional]

17 minutes ago, SuperCookie78 said:

I thought AMD finnaly  decided to get a CUDA license. 

No they launched the Boltzmann initiative. Included are easy ways to program OpenCL, convert CUDA code to OpenCL, and use C language instead of CUDA via their HIP tool:

http://www.amd.com/en-us/press-releases/Pages/boltzmann-initiative-2015nov16.aspx

[SuperCookie78]

Just now, Notional said:

No they launched the Boltzmann initiative. Included are easy ways to program OpenCL, convert CUDA code to OpenCL, and use C language instead of CUDA via their HIP tool:

http://www.amd.com/en-us/press-releases/Pages/boltzmann-initiative-2015nov16.aspx

Oh ok thanks for clearing that up

[Doobeedoo]

So they could potentially make the largest die ever huh, if so that would be impressive. With all the new node and architecture I can't wait to see what they will come up with.

[RagnarokDel]

22 minutes ago, SuperCookie78 said:

I thought AMD finnaly  decided to get a CUDA license. 

they have a cuda license. They've had it for a while actually. (months)

[Citadelen]

I want AMD to make real progress with power consumption, and not just rip out a feature and call it a day. (looking at you Maxwell)

[Watermelon Guy]

There saying this is equal in density to 450mm in 28nm

[patrickjp93]

7 minutes ago, Watermelon Guy said:

There saying this is equal in density to 450mm in 28nm

No. Every node change doubles density. 32-22 doubled Intel's, and 22-14 doubled it again. 28-20 came close to doubling density for TSMC and Samsung. 20-14/16 is another such case. Dropping from 28 to 14 should nearly quadruple density. The transistors should be ~1/4 the size on average.

[-BirdiE-]

4 hours ago, byalexandr said:

Just imagine.... Polaris cards more powerful than the R9 Nano but just as short and half height....

Implying 100% of a card's size is related to the manufacturing size...

 

That's why graphics cards have been getting way smaller over the years.... OH wait..

[patrickjp93]

46 minutes ago, -BirdiE- said:

Implying 100% of a card's size is related to the manufacturing size...

 

That's why graphics cards have been getting way smaller over the years.... OH wait..

Well yes. We've only gotten smaller since the VooDoo cards. God do you people pay attention? /s