• 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.

Radeon RX 480 "Polaris" Launched at $199

All I know is that my 290@1050 gets 7.1 in VR bench :p So something just doesn't add up. For some reason RX480 is rubbish in this benchmark.

it is not that it is rubbish, it is that we dont have the talke from the presentation that go with the slide. It can very well be that the score it has was for it as base clocks in a low power mode. considering 'Low Power' was in big letter above the graph on that slide. it could have very well be talking about the mobile vr backpack computer.
 
it is not that it is rubbish, it is that we dont have the talke from the presentation that go with the slide. It can very well be that the score it has was for it as base clocks in a low power mode. considering 'Low Power' was in big letter above the graph on that slide. it could have very well be talking about the mobile vr backpack computer.

You can debate no more because the VR figures are confirmed now.
http://www.techpowerup.com/forums/t...leaked-and-tested.223351/page-22#post-3479235
 
You can debate no more because the VR figures are confirmed now.
http://www.techpowerup.com/forums/t...leaked-and-tested.223351/page-22#post-3479235

Exactly. Which makes no sense. There are also some other benchmarks floating around against custom overclocked 970:

Doom
Medios (base 380: 55-60fps):
RX 480: 163%
GTX 970: 147%
R9 380:100%

Mínimos (base 380: 40-45fps):
RX 480:158%
R9 380:100%
GTX 970: 93%

Firestrike
Graphics Score (base de la R9 380: 8700-8800 ptos):
RX 480: 149%
GTX 970: 142%
R9 380: 100%


Gears of War: Ultimate
Medios (base 380: 55-60 fps):
RX 480: 158%
GTX 970: 153%
R9 380: 100%

Mínimos (base 380: 50-55fps):
RX 480: 119%
GTX 970: 115%
R9 380: 100%
 
Yet the performance in games shows it punching above that weight. It could more than likely be an issue with OpenGL then and some functionality of the card not being fully enabled with it yet.

Or you know the Steam VR test being wonky as hell; where the GTX 580 scores around 10 points.
 
I think it is a little oposite of what you are saying.

GCN is much more brute-force, huge compute potential but much less transistors used to fully exploit that compute performance. AMD tried with Async compute and spent considerable transistors on hardware based scheduling but without fixing the actual bottlenecks in the GPU design.

Kepler, and especially Maxwell and pascal have a lot more finesse. They have less theoretical performance because instead of just throwing more and more compute units at the GPU they dedicate more and more transistor budget to making the actual compute resources fully utilized and remove all kinds of bottlenecks, including DX11 limitations. Therefore, there is far less to be gained form async compute or DX12 because the hardware is not limited to the same extent.
This is pretty much spot-on. For the most part.

GCN has higher brute-force potential. But they definitely make sacrifices to get there, and it affects them. Combined with reknowned lackluster drivers that take 6-18 months to actually come good in DX11, you have the perfect example of one company living in the present and making the most of it, and the other thinking arguably too far ahead and not getting the most of their hardware until quite late(at which point there are new and better cards available).

All in all, I'd say the trend is that AMD has been a better bet if you're not the type to upgrade every 1-3 years, but otherwise Nvidia seems to invest more of their design and resources into the here and now, and thus while not as great if you want a card to last you many years, will serve you very well from Day 1 up until you upgrade.

Of course AMD are making bigger strides nowadays with drivers. So it could be that they come out swinging harder on Day 1 and thus dont quite have the legs they used to have relative to Nvidia. People will spout DX12, but there's not going to be half as many DX12 games as some people think in a year's time. We still dont even have any TRUE DX12 games right now. Just DX12 branches, as nobody wants to make a Windows 10-only game just yet.

That said, while Nvidia have many DX12 features available to them, more than some will give them credit for(there's more to DX12 than async compute), the aformentioned advantage with better driver support for Nvidia in DX11 goes away in DX12. That doesn't necessarily make AMD cards inherently better for DX12, what it means is that Nvidia's advantage is negated and it comes more down to how the low level code is programmed. It's why we've seen several examples of DX12 backends in games not exactly help AMD cards to any meaningful degree, but still hurt Nvidia cards some(with Maxwell at least).

DX12 is going to be a *very* complicated situation going forward. It is not half as straightforward as 'AMD=better, Nvidia=bad' as some people have been sold. It's its own new can of worms that we're really only just starting to peel open right now. And so long as Nvidia and AMD provide different GPU architectures with their own nuances, there is going to be disparity in low-level optimizations and how they correlate to performance for certain cards.

Honestly, I think it could easily be a huge mess and if it becomes the norm, we could head back into the days where there was a much larger disparity between how a game ran depending on what brand card you had. Despite still having differences the last 5-8 years or so, I dont think people realized how good we've had it that optimizations generally meant fairly predictable and comparable performance whatever you bought.
 
I'll be holding out for Msi etc before buying. Even though I don't overclock much myself am very interested in seeing how much extra performance they get out of these.
 

The hype went down for the last few days but looking good again.

Some new firestrike results:

480@1266MHz - gfx score = 12779
480@ 1330MHz - gfx score = 13403

http://videocardz.com/61475/amd-radeon-rx-480-rumors-part-7

Seems to be around 390x/Fury nano level. AIB versions with higher overclock possibility should make this a nice card indeed.
 
Last edited:
I think it is a little oposite of what you are saying.

GCN is much more brute-force, huge compute potential but much less transistors used to fully exploit that compute performance. AMD tried with Async compute and spent considerable transistors on hardware based scheduling but without fixing the actual bottlenecks in the GPU design.

Kepler, and especially Maxwell and pascal have a lot more finesse. They have less theoretical performance because instead of just throwing more and more compute units at the GPU they dedicate more and more transistor budget to making the actual compute resources fully utilized and remove all kinds of bottlenecks, including DX11 limitations. Therefore, there is far less to be gained form async compute or DX12 because the hardware is not limited to the same extent.

Actual the opposite is true . Nvidia is brute foce and AMD is more finesse and optimisation hence great deal to be gained from driver improvements.
 
We still dont even have any TRUE DX12 games right now. Just DX12 branches, as nobody wants to make a Windows 10-only game just yet.

Not disagreeing with you as you make a few good points in your post but Quantum Break PC is DX12 Win 10 only. It apparently didn't sell too well but AMD put in a good show with 390X hot on the heels of 980Ti in that game (as we've seen in other DX12 "branches").
 
Actual the opposite is true . Nvidia is brute foce and AMD is more finesse and optimisation hence great deal to be gained from driver improvements.
Nope. You're confusing the long-term driver improvements from AMD as 'finesse' when it's really just poor Day 1 drivers, leaving more room for improvement later.

AMD's lackluster DX11 drivers have been well demonstrated by this point.

Not disagreeing with you as you make a few good points in your post but Quantum Break PC is DX12 Win 10 only. It apparently didn't sell too well but AMD put in a good show with 390X hot on the heels of 980Ti in that game (as we've seen in other DX12 "branches").
Right, I forgot about rushed UWP releases like these.

I figure most UWP releases will skew AMD simply because the whole premise behind the platform is not to spend much extra time optimizing for PC in general. They take console code and port it over for the most part.

Should be pointed out that like many DX12 titles so far, it suffered from quite a few issues.
 
Last edited:
Back
Top Bottom