AMD's next GPU uarch is called "Polaris"

[sontin]

Sure, nVidia has missed the hype train for a product which you can only buy in june.

 

[Techhog]

People already forgot the Fiji "scores" they manipulated? Dont put much into it. Because its certain that reviews will show something different. Marketing never matches reality.

The only ways to really manipulate it here though are to lie about the settings used or to have some trickery going on with vsync. If they haven't done that, then they're accurate... at least for that particular game.

That said, you think that the results are off by 100-150% (as in, you think that the card actually uses 2-2.5x as much power as what's presented here). That's a bit on the extreme side, don't you think?

 

[ShintaiDK]

You make a lot of assumptions on what I think

Can you tell me what exact chip AMD used?

Example:
My GTX980 can also easily beat a GTX950 using half the power. If I can choose the setting.

 

[lilltesaito]

So, showing of working silicon for a product released in 6 months is impressive?

nVidia had never shown Maxwell prior the launch and yet Maxwell was impressive at launch.

It seems like AMD is taking a different approach this time around. Instead of trying to hide everything, they are sharing more. Everyone says that AMD marketing sucks, why get upset about them trying to change that? Seems like it is better then all the slides they would show, now they have a working(unfinished) product they are showing.

I do not understand why people have to come into threads and try to lower hype. Does it some how affect people's personally? Does this hype come to your house and trash it or something?

I think it is perfectly fine correcting wrong information. Good example is when someone thought it was a 970 being tested in the slide.

To answer your question, yes I think it is impressive. Because they have a non finished video card working and showing almost half the power draw of the same level card. We have no clue on what Nvidia has, so no point in bring that up since AMD does have a working card out showing what it can do.

What would be cool to see is Nvidia does the same thing and shows us a working card that is showing us improvements.

I like the new changes AMD has been making, hated that they did not release any information about fury. I think they know that it does not matter if they try to get keep everything hidden because Nvidia still seems to find out about it and gets the jump on them(980ti release).

 

[KaRLiToS]

"Wow"?
Maybe i miss something: AMD is comparing a 16nm FinFet chip to a 12months old chip with nearly 2x times the transistors in a AMD biased game. They archived a 2,5x improvement in efficiency.

nVidia released the GTX980 11 months after the 290X, which has less transistors and archived a 1.9x to 2.0x improvement...
Would have nVidia demonstraded the GTX980 in march 2014 nobody would believed it.

Yet when AMD needs a much longer timeframe, a new process node and nearly twice the transistors, everybody is screaming "Wow".

I guess the hype train is back and now bigger and better than ever.

What the hell are you talking about? If everyone was saying "WOW" like you mentionned, the market share would be in favor of AMD. But it is not the case.

And also, would you prefer AMD to use a Gamewreck (Gimpwork, I mean Gameworks) title instead? And don't forget that Gamewreck develpper program is the new technology that is actually killing the PC gaming industry and bringing pain to all gamers. I think I understand why AMD uses an AMD game title. Duh

 

[Techhog]

You make a lot of assumptions on what I think

Can you tell me what exact chip AMD used?

Example:
My GTX980 can also easily beat a GTX950 using half the power. If I can choose the setting.

You said that you expect 2x or less efficiency compared to GCN 1.1, so if myt "assumption" is incorrect then you need to get your thoughts straight.

Oh cool, you can back this up, then!

 

[96Firebird]

...hated that they did not release any information about fury.

Wat.

 

[Tuna-Fish]

Well, if this is true it means that GPU with performance levels at 1080p between R9 285 and 280X will use 30W.

WTF?

You have to remember the usual disclaimer for these kinds of demonstrations: It's very unlikely that they lie, but they get to pick the load. You can bet your ass they chose the exact setup of game/settings/system that shows their new stuff in the best possible light. This is not average improvement, this is the very best they could find after looking through all possible options. For example, while the CPU they used can draw a lot of power, do we know how much it draws running the program they used to demonstrate?

That in mind, finfets are a huge improvement. In a lot of ways that are critical to GPUs, we get the equivalent of two node shrinks in one.

 

[Techhog]

Wat.

We got almost nothing about Fiji beyond HBM talk before the full reveal, to be fair.

 

[ShintaiDK]

You said that you expect 2x or less efficiency compared to GCN 1.1, so if myt "assumption" is incorrect then you need to get your thoughts straight.

Oh cool, you can back this up, then!

I never said GCN 1.1 did I? And I never said less.

AMD expects 2x performance/watt over GCN 1.2 when they actually have to talk to people with money.

Only a complete fool would trust a PR presentation with unknown factors and setup. All running specific tweaked setups. In this case 1 single game. In a game where a 110W TDP 370 GCN 1.0 equals a 90W TDP GTX950. Now move to GCN 1.2 and do the node shrink. Even better if a higher class chip is being used for 14nm. PR is and will always be PR.

 

[lilltesaito]

Wat.

I guess I should say sooner, only thing people were going by were leaks that had no real backing.

 

[beginner99]

A tiny chip. what a surprise...NOT. Was obvious that the will first release small die due to yields and mobile.

Till we actually see something that beats a 980TI on performance or beats used 290/290x (or new one bought about 12 month ago) on price/performance it will be 2017.

 

[Head1985]

A tiny chip. what a surprise...NOT. Was obvious that the will first release small die due to yields and mobile.

Till we actually see something that beats a 980TI on performance or beats used 290/290x (or new one bought about 12 month ago) on price/performance it will be 2017.

Nope..I am pretty sure pascal/GTX970/980 successor will beat TITANX around 40%.
If AMD want be competetive they need GPu around 300-350mm2 on 14/16nm

 

[Techhog]

I never said GCN 1.1 did I? And I never said less.

In AMDs case you may end up seeing 390 performance in a 125-150W power area.

275/2 = 137.5W, and that's when you include the high-clocked GDDR5 in that calculation. Compare it to the 290X, which offers similar performance at a 250W TDP, and the 125W becomes the minimum. Knowing how pessimistic you are (about all but one primary color), you certainly feel that the higher end of the spectrum is more likely than the lower end.

 

[Kenmitch]

I never said GCN 1.1 did I? And I never said less.

AMD expects 2x performance/watt over GCN 1.2 when they actually have to talk to people with money.

Only a complete fool would trust a PR presentation with unknown factors and setup. All running specific tweaked setups. In this case 1 single game. In a game where a 110W TDP 370 GCN 1.0 equals a 90W TDP GTX950. Now move to GCN 1.2 and do the node shrink. Even better if a higher class chip is being used for 14nm. PR is and will always be PR.

I guess you would have been happier with just slides, and maybe a prototype card held up being held together with wood screws then.

It's too early for the AMD will never do good in your eyes smear campaign.

 

[csbin]

http://www.hardware.fr/news/14456/amd-annonce-polaris-future-architecture-gpu.html

Il est équipé de mémoire GDDR5. Nous n'avons pas encore de nom pour ce GPU et n'avons pu l'apercevoir que brièvement sans pouvoir en prendre une photo. Tout juste de quoi apercevoir qu'il s'agit effectivement d'une petite puce et d'un packaging compact.

 

[PhonakV30]

I never said GCN 1.1 did I? And I never said less.

AMD expects 2x performance/watt over GCN 1.2 when they actually have to talk to people with money.

Only a complete fool would trust a PR presentation with unknown factors and setup. All running specific tweaked setups. In this case 1 single game. In a game where a 110W TDP 370 GCN 1.0 equals a 90W TDP GTX950. Now move to GCN 1.2 and do the node shrink. Even better if a higher class chip is being used for 14nm. PR is and will always be PR.

Where did they say "over GCN 2" ?

 

[Techhog]

http://www.hardware.fr/news/14456/amd-annonce-polaris-future-architecture-gpu.html

Il est équipé de mémoire GDDR5. Nous n'avons pas encore de nom pour ce GPU et n'avons pu l'apercevoir que brièvement sans pouvoir en prendre une photo. Tout juste de quoi apercevoir qu'il s'agit effectivement d'une petite puce et d'un packaging compact.

Well, there goes the "underclocked higher-end chip" theory

 

[Techhog]

Where did they say "over GCN 2" ?

FYI, AMD would call GCN 1.1 "GCN 2"

 

[sm625]

This image has been going around. Do you guys remember the name of the technique that Nvidia uses to save power that is wasted processing frames that are not needed? I assume that when a game is VSYNC capped at 60 fps, it is saving power vs running at 150fps or whatever max the hardware is capable of hitting. Are there any benchmarks that compare 60fps capped power consumption vs uncapped on a 950? If we had that info we would be able to better assess just how good Polaris is.

 

[Headfoot]

Did any of you actually read the Anandtech article???

They specifically called out that it was capped at 60 fps. AMD also similarly in the last 6 months released a new driver with frame rate capping capabilities, and boasted about how it increased perf/watt. They picked a 950, not a 960, so that more of the chip would have to be spun up to higher frequencies to hit 60 (wider w/ lower clock = more power efficient, narrower with higher clock = less power efficient). The new chip can likely do more than 60fps in SWBF, and the 950 is probably nearing its cap, and this was intentionally chosen. Do the math.

Those numbers they showed are certainly true, but they've also picked the precisely best light to show it in. As anyone should expect from a products company marketing its product...

 

[Kenmitch]

Did any of you actually read the Anandtech article???

They specifically called out that it was capped at 60 fps. AMD also similarly in the last 60 months released a new driver with frame rate capping capabilities, and boasted about how it increased perf/watt. They picked a 950, not a 960, so that more of the chip would have to be spun up to higher frequencies to hit (wider w/ lower clock = more power efficient, narrower with higher clock = less power efficient). Do the math.

Those numbers they showed are certainly true, but they've also picked the precisely best light to show it in. As anyone should expect from a products company marketing its product...

....Or they just picked the current NVidia offering that is currently in the price point of this Polaris chip.

 

[ShintaiDK]

Well, there goes the "underclocked higher-end chip" theory

From what information? That its a GDDR5 part?

 

[lilltesaito]

They specifically called out that it was capped at 60 fps. AMD also similarly in the last 60 months released a new driver with frame rate capping capabilities, and boasted about how it increased perf/watt.

I have no doubt that is why they used the 60fps cap.

 

[Headfoot]

....Or they just picked the current NVidia offering that is currently in the price point of this Polaris chip.

That's a given. I doubt pricing is yet firm anyways