gamegpuKiller Instinct DX12 Benchmarks980TI vs 290X)

[xthetenth]

He guessed, you guessed. Yours is no better than his.

Simple as that. :thumbsup:

Yeah, it's not like one of the guessers is working from hugely worse information than the other. I mean that google translate is exactly the same as the words as originally written in the context of the language they're written.

I'm so glad we live in the glorious future where translation is a solved problem.

 

[AtenRa]

I really want to see prices of Kepler and Maxwell cards on the used market at the end of 2016. If every DX-12 game in 2016 is like the ones we seen so far, R9 290/X and R9 390/X may sell at the same price as GTX 980Ti.

 

[guskline]

Now I feel bad for selling 3 x R9 290x.




No, why would I say that, my last R9 290x is as good as a GTX 980ti.

Not if you use the money for a Big Vega next year.

 

[tential]

isnt the lesson you pay twice as much for a card that performs as well as the mid range?
I mean, seriously something is seriously wrong with the nvidia approach.

People don't look at benchmarks. An acquaintance I know has a gtx 970 and he sold his r9 290 to get it. His friends have Nvidia, Nvidia is good, Nvidia drivers are good, so gtx 970!

He has had to format 3+ times since, has worse performance in the games he plays, and he does not usually use ultra textures/settings....

It doesn't matter if you have performance or longevity. It matters that people believe you do, and that is what counts.

Nvidia makes people believe and that's something that will just never show in the numbers.

 

[Glo.]

I can agree with you on that topic, tential. I am experiencing constant attacks on my person on another technological forum, for talking about objective facts, that AMD is not worse than Nvidia, when I bring up understanding to the discussion. Only Nvidia is the best, and AMD is sh**.

 

[Kuiva maa]

That's not at all what it says. The Editor states that due to circumstances related to AMD, Fury-based videocards will be removed/absent from future testing for several weeks. The most logical conclusion is that GameGPU received hardware review samples directly from AMD (hence why they had no Fury cards in the charts for months after launch), and now AMD wants to rotate these cards by giving them to another reviewer in Russia. It also explains the lack of i5 6600K, i7 6700K, R9 390, 390X, etc. since these products were never sent to them.

If you ask me it is kinda poor decision from AMD to do so. Rotating press samples is perfectly normal but gamegpu.ru ,regardless of being good or bad at what it does, is having great reach among tech forum members and opinion leaders. If I was AMD i would be sending them one unit of each significant new GPU.

 

[thesmokingman]

If you ask me it is kinda poor decision from AMD to do so. Rotating press samples is perfectly normal but gamegpu.ru ,regardless of being good or bad at what it does, is having great reach among tech forum members and opinion leaders. If I was AMD i would be sending them one unit of each significant new GPU.

I don't think they should show favoritism. That potentially could lead to quite the slippery slope. Reviewers should be buying their own hw if we are to get an impartial press. And the extreme of that is hardforum, who got a chip on their shoulder because they expected to receive X card and didn't get one. We talk about our youth feeling entitled, but look what happens with reviewers.

 

[KaRLiToS]

Not if you use the money for a Big Vega next year.

I agree with you. But I won't buy a big Vega. I'll buy 2 or more. :thumbsup:

Depends of what Pascal looks like and how it is priced though.

 

[AbRASiON]

This is happening to more and more games recently. lol at people who bought $600 980ti cards, getting same performance as a 3 year old $200 card.

Dumb question:
What's to stop nvidia fixing their DX12 drivers to keep up?

 

[Silverforce11]

Dumb question:
What's to stop nvidia fixing their DX12 drivers to keep up?

Their hardware.

We are all waiting to see if Pascal will be capable.

 

[RussianSensation]

Dumb question:
What's to stop nvidia fixing their DX12 drivers to keep up?

Unlike overtessallation that can be reduced, you cannot fix something that's fundamentally absent in the hardware. Card A can do tasks X+Y faster than card B can do X, then Y.

Another way to look at it is with parallel Async Compute you can do something 10-30% faster than using a serial method of waiting for A to do B then waiting for B to do C. That's why Async compute is not at all like tessellation. Overtessallation reduces performance on all hardware. Excessive compute improves performance on architectures that can do it, and penalizes architectures that cannot since they can't take advantage of parallel computation. Async compute is like a person having 3 arms (AMD) vs. one with 2 arms (NV). You give them 3 jobs to do. The 3rd arm is only allowed to perform specific tasks, but if you don't have it, you will never speed up the work since neither of your 2 arms can handle the specific tasks of a 3rd arm at the same time while the 2 arms are busy -- that's NV right now in DX12.

In simplest terms, Async Compute is like multi-cores for CPUs - X cores run audio, Y cores run AI, Z cores run graphics at the same time. NV can't split compute and graphics workloads without a huge performance penalty, making it a worthless feature. It basically means NV can do compute work loads really well (Tesla) and graphics workloads really well (GeForce), but it cannot do both concurrently well -- AMD's GCN can under DX12 or OpenCL.

All that extra power consumption from is partly because GCN has so many more functional units, more advanced command processor, ACE engines, more shaders, etc.

Having said that, some of these DX12 games don't even appear to use much Async Compute, which means non-GCN architectures suffer since the developers probably don't want to spend extra $ optimizing beyond console architecture. This has been called the 'console effect'. I believe NV tried to combat this with GameWorks but they don't have unlimited funds to buy out every AAA game.

 

[Pantalaimon]

Predicting the future is a VERY dangerous thing to do

That's a weird thing to say. Are you saying businesses should not try to predict the future?

 

[TheELF]

That's a weird thing to say. Are you saying businesses should not try to predict the future?

Are you saying businesses haven't fallen flat on their faces trying to predict the future?

 

[3DVagabond]

Are you saying businesses haven't fallen flat on their faces trying to predict the future?

How would you suggest they design chips for five years from now?

 

[ThatBuzzkiller]

That's a weird thing to say. Are you saying businesses should not try to predict the future?

I'm not saying they shouldn't but in an industry with short product cycles such as chip design, software and semiconductors, long term outlook isn't as important compared to the short term ...

When you sacrifice current performance for future performance, it is a risky game and you don't know whether or not it will pay off since there are no guarantees that future software will be designed in mind with that hardware ...

 

[airfathaaaaa]

And what happens if the future isn't the same as you predicted ? It becomes a waste just like OpenGL ...

OpenGL was once forward looking but you don't see many games using it because of the mistakes that compounded in it's design ...

Predicting the future is a VERY dangerous thing to do and look where it's got AMD in their market share predicament, they should learn to capitalize more aggressively on what they have currently ...

Had Killer Instinct been a DX12 game, AMD would have had another Quantum Break style domination on their hands a la 390 = 980Ti OC ...

well they didnt really predicted the future
they have won all of the consoles literally everyone that wanted to develop a game would have done on an amd arch eventually pushing for an api that would provide a homogeneous enviroment thus leading to a massive save of money for devs was the only way to go

 

[ThatBuzzkiller]

well they didnt really predicted the future
they have won all of the consoles literally everyone that wanted to develop a game would have done on an amd arch eventually pushing for an api that would provide a homogeneous enviroment thus leading to a massive save of money for devs was the only way to go

Well they did, even if practically all the AAA developers were forced to optimize for consoles AMD still weren't sure that it would translate to PC hence their initiative for Mantle ...

It took almost 2 years for DX12 to materialize in a bare bones form and it will probably take another year for mass adoption too ...

AMD got very lucky this year but they depend too much on external factors for success ...

 

[Pantalaimon]

Are you saying businesses haven't fallen flat on their faces trying to predict the future?

No. His post just stressed the point which to me sounded like he is against it. Predicting the future always has risks, but businesses have to do it to whatever degree they can.

 

[Adored]

Well they did, even if practically all the AAA developers were forced to optimize for consoles AMD still weren't sure that it would translate to PC hence their initiative for Mantle ...

It took almost 2 years for DX12 to materialize in a bare bones form and it will probably take another year for mass adoption too ...

AMD got very lucky this year but they depend too much on external factors for success ...

It's not luck I can assure you. It's a carefully executed master plan and you'll soon see just how incredible it is.

 

[Silverforce11]

It's not luck I can assure you. It's a carefully executed master plan and you'll soon see just how incredible it is.

You know, I didn't think small Polaris would be competitive vs GP104, but damn, these new 2016 game lineup...

I mean if sites actually use these new DX12 titles in their updated benches, it's instant GCN domination.

 

[poofyhairguy]

AMD got very lucky this year but they depend too much on external factors for success ...

I would say the whole PC gaming market is dependant on external factors. Most of the games are console ports whose quality for either type of GPU depends on the resources thrown at the port. The future of many Directx 12 games will depend on what direction Microsoft wants to go in gaming in relation to the Xbox.

VR for now is mostly a PC only experience, but the rest depends on external factors. That is why Nvidia is chasing compute and AMD needs console wins.

 

[VulgarDisplay]

Are you saying businesses haven't fallen flat on their faces trying to predict the future?

I think the important thing for AMD which is bad for nvidia right now is that they stopped predicting the future and started dictating it by winning all the console designs.

Sent from my XT1575 using Tapatalk

 

[ThatBuzzkiller]

It's not luck I can assure you. It's a carefully executed master plan and you'll soon see just how incredible it is.

This so called "incredible" plan got AMD into the mess their in with an 80:20 split in favour of Nvidia ...

I don't think you understand how easy it is to derail such a plan especially in the situation AMD is in ...

Polaris 10 needs to hammer the GP104 at DX12 in the next 6 months after it's release to deem AMD's plan as success worthy if you can even call it that ...

 

[ThatBuzzkiller]

I would say the whole PC gaming market is dependant on external factors. Most of the games are console ports whose quality for either type of GPU depends on the resources thrown at the port. The future of many Directx 12 games will depend on what direction Microsoft wants to go in gaming in relation to the Xbox.

VR for now is mostly a PC only experience, but the rest depends on external factors. That is why Nvidia is chasing compute and AMD needs console wins.

Maybe but AMD is literally at the mercy of everyone ...

They need to hope that Pascal is a dud in DX12 the same way Maxwell is before we can start evaluating how realistic their chances are ...

 

[book_ed]

This so called "incredible" plan got AMD into the mess their in with an 80:20 split in favour of Nvidia ...

I'd say their mistake was to originally release the R290/x with those terrible coolers. It left a bad rep. R290 was their gtx970 and then some more, in the long term. R290 was about $400 and nvidia had the 780ti at $700. Performance difference was about 14% at their launch date, so basically when nVIDIA was at its best.

Also when gtx970 was at its best (when it launched), in 1440p was 4% faster and 2% faster in 4k compared to a stock R290. This was by all means an incredible performance to be able to keep up with 2 generations. However, if you put emphasis and look only at 1080p in those times...

Of course, it went with AMD bad, nVIDIA good and power consumption above all lead to a situation in sales that doesn't actually paints a true to life picture. If the situation was in reverse, I very much doubt AMD would have had the same sales and market advantage over nVIDIA, just like when it had the upper hand against Intel, but still couldn't put that into a significant sale number.

At the end of the day it's all about user perception. AMD left way to much for the cards to speak for themselves when they should have been more careful about the state of the hardware in which is released and their relations with game developers. Polaris/Vega could be great, but I hope they don't make the same mistakes over and over again - because R3xx/Fury series was another time in which they failed a bit.