• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Polaris architecture – GCN 4.0

SiDeards73 said:
Ok but did AMD not say "more than 2x the performance" infact their slide states 2.5x performance per watt

https://www.extremetech.com/wp-content/uploads/2016/03/AMDGPU.jpg

So where does that put us?
Click to expand...

Vs Fury-X its this [-------------------------] much
Vs Fury-Nano its this [-----------------------------------------------------] much
Vs 290X its only this [-----------------] much

Take your pick.








































The poin't is ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------we don't know. :p
 
SiDeards73 said:
No one knows what it will draw, we can all speculate, maybe it will draw 80w, maybe it will draw 100w, maybe 120w, nobody knows.

Maybe the cards will give between 2.2 and 2.5x perf per watt, maybe they wont, nobody knows.

Maybe the Polaris reveal is just going to be Laptop based gpu's only, maybe its going to be a mix of laptop and discreet, nobody but AMD knows.
Click to expand...

it was mentioned before that P10 will have a maximum TDP of 175W but run far bellow that under normal operation.
 
humbug said:
2.2x the performance per watt, according to AMD, not 2x.

The 290X has a 300 Watt TDP, no official rating, just that, the actual power consumption is around 250 to 275 Watts depending on where you look.

It also depends on what GPU we are talking about, the Fury-Nano, for example is a 175 Watt card and is much faster than the 290X

Even the Fury-X which one would argue is the one they are talking about given its the newest has the same power levels as the 290X, but is again much faster.
Click to expand...

The most recent numbers from AMD which they told investors is 2x.
THE 290X actually pulls more than the stated 275w

AMD explicitly said 2X relative to their existing mainstream, not their existing high-end. And the Fury products are fairly irrelevant because they use HBM memory cutting around 30w.


Look, you can go on fantasizing all you want but AMD are making it very clear what to expect
 
humbug said:
Vs Fury-X its this [-------------------------] much
Vs Fury-Nano its this [-----------------------------------------------------] much
Vs 290X its only this [-----------------] much

Take your pick.

The poin't is ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------we don't know. :p
Click to expand...

This is very true, like i said before, nobody but AMD knows, but there is a lot of misinformation spread on this board, mostly by Nvidia fans like DP, blatantly stating its 2.0x performance, when AMD themselves have stated 2.5x performance, but in a way his point is valid as nobody knows the metric its being measured against.
 
D.P. said:
The most recent numbers from AMD which they told investors is 2x.
THE 290X actually pulls more than the stated 275w

AMD explicitly said 2X relative to their existing mainstream, not their existing high-end. And the Fury products are fairly irrelevant because they use HBM memory cutting around 30w.


Look, you can go on fantasizing all you want but AMD are making it very clear what to expect
Click to expand...

Can you provide a source for this 2x performance please? i would like to see it, as i was not aware they had changed stance from 2.5x
 
SiDeards73 said:
Ok but did AMD not say "more than 2x the performance" infact their slide states 2.5x performance per watt

https://www.extremetech.com/wp-content/uploads/2016/03/AMDGPU.jpg

So where does that put us?
Click to expand...


This is AMD's mo=st recent statement
http://finance.yahoo.com/news/edited-transcript-amd-earnings-conference-051600440.html

Polaris delivers double the performance per watt of our current mainstream offerings
Click to expand...


But even if you start using higher figures it still makes P10 at 390X performance using much more than 80w 2.5X AMD mainstream cards still puts AMD behind Pascal.
 
D.P. said:
The most recent numbers from AMD which they told investors is 2x.
THE 290X actually pulls more than the stated 275w

AMD explicitly said 2X relative to their existing mainstream, not their existing high-end. And the Fury products are fairly irrelevant because they use HBM memory cutting around 30w.


Look, you can go on fantasizing all you want but AMD are making it very clear what to expect
Click to expand...

I thought the R9 390 cards are more enthusiast as defined by their price points??

Mainstream should be the R9 380 and R9 380X based cards.
 
So the official slide says one thing and a statement from Dr Lisa Sue in a conference call says another.

Take your pick time again guys but the official figures have not changed.

what a merry-go-round in here again.....
 
CAT-THE-FIFTH said:
I thought the R9 390 cards are more enthusiast as defined by their price points??

Mainstream should be the R9 380 and R9 380X based cards.
Click to expand...


Yes, 380-380X is probably what AMD are referring to:

https://www.techpowerup.com/reviews/ASUS/R9_380X_Strix/24.html
https://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_1080/27.html

The 1080 is well over twice the efficiency of the 380, under real world testing and not theoretical marketing numbers. The new process will get them about 70% faster for the same power draw. Generational improvements are around 15-20% typically. Maybe AMD are making a very big leap at 20-25% gain. But that just meets Pascal.


Of course there are ways you can increase performance per watt, like downclocking a big chip (Nano), water cooling (FuryX), HBM memory etc.
 
humbug said:
So the official slide says one thing and a statement from Dr Lisa Sue in a conference call says another.

Take your pick time again guys but the official figures have not changed.

what a merry-go-round in here again.....
Click to expand...


The details different because AMD are referring to different cards. ANd one is a pure marketing PR job that doesn't require much legal basis, the other is an important legally binding statement to investors where lieing would cause them a lot of ramifications and future legal cases.
 
D.P. said:
Yes, 380-380X is probably what AMD are referring to:

https://www.techpowerup.com/reviews/ASUS/R9_380X_Strix/24.html
https://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_1080/27.html

The 1080 is well over twice the efficiency of the 380, under real world testing and not theoretical marketing numbers. The new process will get them about 70% faster for the same power draw. Generational improvements are around 15-20% typically. Maybe AMD are making a very big leap at 20-25% gain. But that just meets Pascal.


Of course there are ways you can increase performance per watt, like downclocking a big chip (Nano), water cooling (FuryX), HBM memory etc.
Click to expand...

Option 4? didn't think of option 4, i would have gone for option 1 myself given that option 1 (Fury-X) is newer... wait no, scratch that, option 2 (Fury-Nano) is newer..... or maybe option 5, is there an option 5?

BTW, i read the whole thing now. Dr Lisa Sue was talking about 14nm FinFet, not the Polaris architecture. when you talk up your production node advantage, which is what she is pandering too in 14nm FinFet you don't say 2.5x better perf per watt even if your talking about the GPU on said node as a whole, because that is misinterpreted as your 14nm FinFet node advantage, there-in false.
 
Last edited:
Phixsator said:
Wauw that HardOCP article sounds like a lot of negative attitude and very little journalism focused on the matter that they were not invited and now AMD **** monkey balls. Even if what HARDoCP is correct it just oozes through with the way it is written. Im sure it one of many reason why AMD dont like them anymore. I keep imagening a child on the playground stomping its feet in anger over not being invite to that other kids birthday party.
Click to expand...

They've been an NVIDIA shill site for what seems like ages. Now to the point that AMD refuse to even acknowledge them - a sad but necessary thing in my view.

It's a shame. Back in the day they used to be one of the very best.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom