Quantum Break Benchmarks, GTX 970 & Radeon R9 390 - PC TURF WARS (Also the shitbox 1)

[-BirdiE-]

23 hours ago, Starelementpoke said:

@-BirdiE- Nvidia does worse in Hitman?

 

7 hours ago, Prysin said:

actually, its not. It loses to the Fury X, Fury, 390X

 

  Reveal hidden contents

 

 

 

 

 

Although, i understand the confusion, as another review site showed Nvidia clearly in the lead under DX12, which is odd, since other DX12 titles show that Nvidia gets teh same or less performance due to ASYNC compute, whilst AMD cards always surge ahead once compute is being used.

 

 

And here is the 1080p one

 

 

Upon closer inspection I see it's the R9 Fury, NOT the Fury X. But to say the 980ti is not ahead of the 390x and 390 would be inaccurate.

 

10 hours ago, ivan134 said:

If its not winning in dx 11 why would it win in dx 12?

Of course. If it performed one way in DX11, why wouldn't in perform the same in a completely different API?

[ivan134]

4 minutes ago, -BirdiE- said:

 

 

And here is the 1080p one

 

 

Upon closer inspection I see it's the R9 Fury, NOT the Fury X. But to say the 980ti is not ahead of the 390x and 390 would be inaccurate.

 

Of course. If it performed one way in DX11, why wouldn't in perform the same in a completely different API?

Hitman uses async compute in dx 12. You know, the thing Nvidia sucks at.

[Prysin]

27 minutes ago, -BirdiE- said:

 

snip

 

Upon closer inspection I see it's the R9 Fury, NOT the Fury X. But to say the 980ti is not ahead of the 390x and 390 would be inaccurate.

 

Of course. If it performed one way in DX11, why wouldn't in perform the same in a completely different API?

Why it would perform diffrently in DX12?

I dunno, because the game engine is built to render certain parts of the game with ASYNC compute, and Nvidia driver has ASYNC disabled, and needs to translate the effect that was supposed to be done with ASYNC compute and use either a proprietary or emulated solution to replicate the effect. Or worse, use OLD solutions that are more resource intensive (DX is after all backwards compatible).

 

A good analogy would be comparing a Koenigsegg Agera and a Lamborghini Aventador. They produce X number of horsepower on regular 95 oct. petrol, but the Koenigsegg produces 150-200 BHP more when using Bio-ethanol.

Now, Why does the same V8 Twin Turbo engine (AMD GPU) produce more BHP when using Bio-ethanol (DX12) then when using 95-oct petrol (DX11?)

Because the engine itself is designed to take advantage of the higher compression rate of Bio-ethanol (DX 12). Higher compression = bigger bang = more power.

The Lamborghini V12 (Nvidia GPU) is meant to run on 95-oct petrol (DX 11). The ECU, injectors and inlet-valves isnt programmed/set up to take advantage of the higher compression rate. This means that while you COULD put bio-ethanol in it, and it would run, it would most likely lose a bit of power, because it will apply the wrong amount air-fuel mix to get the ideal bang. The result is equal or lower power due to the conditions have changed.

 

This is the simplest way i can explain why without getting overly technical on the GPU side and blaming one side or the other for bad planning.

 

Also, here is another source going against the one you linked:

 

The link to the article:

http://www.tweaktown.com/guides/7634/hitman-pc-performance-analysis-directx-12-finest/index2.html

 

 

I am not discrediting PCGamesHardware.de, however i find it funny that there now is TWO sources saying the opposite. This could be a coincidence. But i think that Nvidia card was boosting or boosted manually WAY over its normal advertised clocks, and the AMD cards were probably left at stock.
Why do i think this?

Because of the GTX 980 scores. Look at them, the delta between the 980 and 980Ti is way off the charts. It should be around 30%, but PCGH.de shows a 40-50% delta between the 980 and 980Ti. That is not normal. Not when you consider that both are clocked around 1300MHz from the start. Sure they will boost differently, but the disparity is a bit off.

 

 

[-BirdiE-]

6 minutes ago, Prysin said:

Why it would perform diffrently in DX12?

I dunno, because the game engine is built to render certain parts of the game with ASYNC compute, and Nvidia driver has ASYNC disabled, and needs to translate the effect that was supposed to be done with ASYNC compute and use either a proprietary or emulated solution to replicate the effect. Or worse, use OLD solutions that are more resource intensive (DX is after all backwards compatible).

 

  Reveal hidden contents

 

 

The link to the article:

http://www.tweaktown.com/guides/7634/hitman-pc-performance-analysis-directx-12-finest/index2.html

 

 

 

I am not discrediting PCGamesHardware.de, however i find it funny that there now is TWO sources saying the opposite. This could be a coincidence. But i think that Nvidia card was boosting or boosted manually WAY over its normal advertised clocks, and the AMD cards were probably left at stock.
Why do i think this?

Because of the GTX 980 scores. Look at them, the delta between the 980 and 980Ti is way off the charts. It should be around 30%, but PCGH.de shows a 40-50% delta between the 980 and 980Ti. That is not normal. Not when you consider that both are clocked around 1300MHz from the start. Sure they will boost differently, but the disparity is a bit off.

 

 

 

That comment was sarcasm aimed at someone who was implying that you could know the outcome of DX12 benchmarks based on DX11 benchmarks.

 

Forgot to /s

 

Also, in the benchmarks I saw the AMD cards were winning at 4K as well. I just don't normally like to include 4K benchmarks because, realistically, no one can actually play these games at 4K Ultra if they want decent framerates.