Quantum Break: More like Quantum Broken.

[IllogicalGlory]

Fable was already dismissed as not being async compute enough. Only the AMD sponsored Oxide is good enough as the sole source of information it seems.

We can try see how the game is set when DX12 releases comes out. Quantum Break will be the first full game. Then we can see if putting all the eggs in the same basket was a good idea or not.

I wonder what the next "full DX12 game" will be?

 

[Silverforce11]

Digitalfoundry benchmarked R9 390/GTX 970. R9 390 kills GTX 970 in gameplay benchmarking with perf increase nearing 50-60%. GCN truly is built for DX12/Vulkan and modern APIs. 970 is playing like a card of a lower tier compared to R9 390.

https://www.youtube.com/watch?v=zK2BUeYqLVI

This game has a bizarre 50 fps cap. Not even 60 fps to allow matching vsync refresh. the developers truly lost it.

Yeah was posted yesterday.

We are speculating as to the why is it so slow. NV claims Game Ready on their website and driver. My theory is Async Compute is in play. Remedy will have to use it on Xbone for all their fancy compute based effects. Because UWP DX12 runs on Xbone and PC, its essentially similar code base. That's what is likely going on, they don't even have an exit button in the PC version, cos it's not actually a PC version, it's more an UWP DX12 version that runs on both platforms.

 

[ThatBuzzkiller]

Finally, took AMD long enough to show the fruits of DX12/Mantle so hopefully they have a long enough relative monopoly on it too so they can get back some of the market share ...

 

[raghu78]

http://www.pcgameshardware.de/Quantum-Break-Spiel-15745/Specials/Technik-Test-1191358/

The game needs work as the random game crashes point to a game not ready for launch. Scaling between 980 Ti OC and 980 OC is miserable. Fury OC easily beats 980 Ti OC with 18% higher avg fps and 30% higher min fps.

 

[Silverforce11]

 

[96Firebird]

Edit - There is a note about their graphs at the bottom of the page but I'm having a hard time understanding the translated text...

 

[Silverforce11]

Edit - There is a note about their graphs at the bottom of the page but I'm having a hard time understanding the translated text...

They just put in the avg/min charts. The frame time was done first, it's in the OP.

Basically the frame time is misleading, because it doesn't stutter on AMD in actual recorded gameplay.

https://www.youtube.com/watch?v=zK2BUeYqLVI

But PCgameshardware.de recorded gameplay on a 980 and that stutters a lot. As I said in a few posts, I recant the OP calling it broken, it's only broken on NV actually.

 

[DiogoDX]

Gears launched broken on AMD and now Quantum Break crippled on Nvidia. Realy poor showing for MS so far.

 

[mohit9206]

Well another terrible UWP port.M$ can take Condom Break and shove it up their asses.People shoudn't be buying and supporting such broken games and closed wall platform but then again most gamers have the memory of a goldfish.

 

[airfathaaaaa]

The DX12 era belongs to GCN. Anything not GCN-like is going to take a dirt nap.

There's no trickling down from Pascal, it's arrangement is now GCN-like. It won't be as gimped in games that are optimized for GCN.

The side effect, if NV optimizes it's drivers and GameWorks for Pascal, it's going to neuter Maxwell and especially Kepler (but it's old then, who cares right?).

nvm delete it

 

[ShintaiDK]

Digitalfoundry benchmarked R9 390/GTX 970. R9 390 kills GTX 970 in gameplay benchmarking with perf increase nearing 50-60%. GCN truly is built for DX12/Vulkan and modern APIs. 970 is playing like a card of a lower tier compared to R9 390.

https://www.youtube.com/watch?v=zK2BUeYqLVI

This game has a bizarre 50 fps cap. Not even 60 fps to allow matching vsync refresh. the developers truly lost it.

Seems like its a bug.

" While the 390 is the faster card running that game - there is a problem there with VSync where the 970 is just missing the multiplier and the 390 is just hitting it - resulting in a massive but artificial performance differential"

 

[ShintaiDK]

I wonder what the next "full DX12 game" will be?

I think some CFOs are looking deep into the eyes of the developer managers these days. Asking if this DX12 is need to have or nice to have after the huge disasters

 

[ShintaiDK]

Well another terrible UWP port.M$ can take Condom Break and shove it up their asses.People shoudn't be buying and supporting such broken games and closed wall platform but then again most gamers have the memory of a goldfish.

MS usually dont do refunds, but they do with Quantum Break. That's how bad it is. Even someone up higher in MS must have thought it was too much crap, despite how unlikely such a scenario is.

 

[sontin]

They offered refunds for Gears of War, too.

 

[S.H.O.D.A.N.]

Seems like its a bug.

Where did the quote come from? What 'multiplier' is it talking about? Because the 970 goes from anywhere around 24 to 35 fps.

Edit: On this very page you can see 390 edging out a 980ti and Fury X manhandling everyone else. That's 'hitting the multiplier' too?

 

[ShintaiDK]

Where did the quote come from? What 'multiplier' is it talking about? Because the 970 goes from anywhere around 24 to 35 fps.

https://forums.overclockers.co.uk/showpost.php?p=29369925&postcount=97
https://forums.overclockers.co.uk/showpost.php?p=29370143&postcount=102

 

[Silverforce11]

Seems like its a bug.

Recanted at the source, that was a random guy, not a dev. Desperado make a mistake. Don't perpetuate mistakes.

QB needs fixes from Remedy that much is certain. It's very broken on NV and the vsync cap is pretty lame for everyone.

And about UWP. I did not know MS has a no refund policy.

That's insane. It's like they don't want PC gamers to even use their store.

 

[ShintaiDK]

Its still a bug and its quite obvious how it works.

You really surprised about the PC (gaming) hate from MS(And Sony)? Think about it, nobody even started up Quantum Break on a PC and tried to close it either at the start or after gaming. There is no exit!

 

[S.H.O.D.A.N.]

https://forums.overclockers.co.uk/showpost.php?p=29369925&postcount=97
https://forums.overclockers.co.uk/showpost.php?p=29370143&postcount=102

That makes zero sense, as both cards show consistent variance in framerate and BOTH tank when volumetric lights are on screen by approximately the same value. If the 970 was 'just missing' some arbitrary threshold, it would not budge from the supposed 30fps cap.

It also does not account for the rest of the results presented, unless the 980ti is being held back by some magical multiplier as well.

Edit: To illustrate, in the first scene shown, the fps hover over the 30/45 range for 970 and 390 respectively. When volumetric lights are involved, it's 25/40. Consistently, over a period of ten or more seconds, so we can put that 'average over more than one second' thing to bed. Where's the multiplier?

 

[TheELF]

Where did the quote come from? What 'multiplier' is it talking about? Because the 970 goes from anywhere around 24 to 35 fps.

Edit: On this very page you can see 390 edging out a 980ti and Fury X manhandling everyone else. That's 'hitting the multiplier' too?

The "multiplier" is double vsync,read up on it if you don't know it,if the game is fixed to run at 60FPS but you only manage 59 you will drop to 30FPS although you are capable of 59FPS, that's a massive drop making it impossible to be sure how much faster one card or the other is.

 

[S.H.O.D.A.N.]

The "multiplier" is double vsync,read up on it if you don't know it

I know what it is and I can see it's not involved. You would too, if you only watch the actual video.

 

[Silverforce11]

The "multiplier" is double vsync,read up on it if you don't know it,if the game is fixed to run at 60FPS but you only manage 59 you will drop to 30FPS although you are capable of 59FPS, that's a massive drop making it impossible to be sure how much faster one card or the other is.

Except it doesn't happen when it's output on the display.

Watch that 60 fps video capture from DF. There is very little stuttering (esp on the AMD 390) card, if it's flipping from 50/60 to 30 as the frame time suggest, it would be crazy stutter.

https://www.youtube.com/watch?v=zK2BUeYqLVI

PresentMon may not reflect reality at this point.

Video capture directly from the signal to the monitor is more reliable.

 

[Det0x]

Fable was already dismissed as not being async compute enough. Only the AMD sponsored Oxide is good enough as the sole source of information it seems.

We can try see how the game is set when DX12 releases comes out. Quantum Break will be the first full game. Then we can see if putting all the eggs in the same basket was a good idea or not.

I wonder what the next "full DX12 game" will be?

Would be fun to hear a response from this original poster

 

[ShintaiDK]

Would be fun to hear a response from this original poster

You would if you read the thread

http://forums.anandtech.com/showpost.php?p=38155037&postcount=262

 

[TheELF]

quantum is the first full Dx12,it displays one frame by calculating 3 frames,it renders at 720,even on pc,and runs everything on the GPU it's hard coded to run at 30FPS,even on PC and even when turning of the 30FPS lock it tops out at about ~50FPS.
A lot of this we already saw with older Dx11 ports,still a lot of people where asking for Dx12 so that ports would become easier,you got your wish.


Sadly I can see this becoming the standard for the foreseeable future,games that will run great on tablets and not even 1.5Ghz CPU cores but will greatly underutilize a PC.