computerbaseAshes of the Singularity Beta1 DirectX 12 Benchmarks

[ShintaiDK]

Dude, you of all people should have learnt from the Mantle saga. There were epic threads in which your predictions are all dud.

Mantle lives on, it's Vulkan, it's in Metal, it's in DX12. Soon all gamers will be running AMD code like I said in 2013.

You vehemently believed Maxwell as perfectly capable of DX12 async compute too, epic threads man! Why don't you spend some effort and ask NV where's their DX12 Async Compute they promised in their drivers?

So rather than saying "what if DX12 isn't good for AMD??"... you should say "what if it's true, DX12 will be great for AMD." Your odds of being correct will be higher.

AMD would also be king of DX11. That turned out quite differently.

Now its DX12. Both saving AMD GPUs and FX CPUs. All based on a single AMD sponsored title that seems to be nothing but a paid benchmark.

How is other DX12 titles looking in terms of performance? Because currently you put all your eggs in a single basket called AOTS.

 

[desprado]

Dude, you of all people should have learnt from the Mantle saga. There were epic threads in which your predictions are all dud.

Mantle lives on, it's Vulkan, it's in Metal, it's in DX12. Soon all gamers will be running AMD code like I said in 2013.

You vehemently believed Maxwell as perfectly capable of DX12 async compute too, epic threads man!

So rather than saying "what if DX12 isn't good for AMD??"... you should say "what if it's true, DX12 will be great for AMD." Your odds of being correct will be higher.

2014 and 2015 AMD games did not sell well.

Example
Battlefield front, Star Wars, Dirt Rally, Hitman 2016.

Compare with Nvidia

MGS V, Witcher 3, Fallout 4, Dying light, Mad Max and Rise of the Tomb Raider. They sold well because they are pro Nvidia game.

I wont buy any game which is totally pro AMD game.

The Divison vs hitman 2016

Division already had 1 million user playing beta from Steam which do not include uplay and other 3rd party.

Hitman 2016 do not even have proper 50K user that played beta.

 

[ShintaiDK]

Exactly that is the reason why this game on fund stage. I mean no developer will ever risk 80% of gamers for AMD users. If they do then they will ended up like Hitman 2016, Dirt Rally and Star Wars which are non selling on PC.

Of course not. This is simple business 101. The best selling GPU(ever) is a GTX 970. If you miss that market you already failed.
http://store.steampowered.com/hwsurvey

 

[Silverforce11]

Exactly that is the reason why this game on fund stage. I mean no developer will ever risk 80% of gamers for AMD users. If they do then they will ended up like Hitman 2016, Dirt Rally and Star Wars which are non selling on PC.

You somehow think that GCN optimized will risk 80% of gamers.

You should do some basic maths and compare Xbone/PS4 and PC.

PC is second class citizen if you haven't noticed. All the big studios push out console builds, then throw PC a bone. Console optimization is #1.

If NV wants their GPUs to run the game well, they are the ones who have to sponsor the PC port.

 

[ShintaiDK]

The excuse is already that Fable doesn't use enough async.

Take a wild guess on how the other DX12 titles will end.

 

[Leadbox]

Exactly that is the reason why this game on fund stage. I mean no developer will ever risk 80% of gamers for AMD users. If they do then they will ended up like Hitman 2016, Dirt Rally and Star Wars which are non selling on PC.

What are you saying here? Are you saying developers should make sure the games run better on nvidia than AMD because the nvidia users will not buy their games? Is this the typical nvidia fanboy mindset? If so, then well..........

 

[ShintaiDK]

What are you saying here? Are you saying developers should make sure the games run better on nvidia than AMD because the nvidia users will not buy their games? Is this the typical nvidia fanboy mindset? If so, then well..........

You are asking developers to try and cripple NVidia performance with 75-80% market share for the purpose of what? To support the underdog who doesn't give them the tools they need anyway so they can look good?

 

[flopper]

You are asking developers to try and cripple NVidia performance with 75-80% market share for the purpose of what? To support the underdog who doesn't give them the tools they need anyway so they can look good?

Last I checked ashes AMD fury destroys nvidia GPUs
just saying

 

[Leadbox]

You are asking developers to try and cripple NVidia performance with 75-80% market share for the purpose of what? To support the underdog who doesn't give them the tools they need anyway?

No, I didn't ask that, you did, my questions are in the post you quoted. So running better on AMD is only because they're crippling nvidia performance right? Using dx12 ASc is trying to cripple nvidia, ok didn't know that.

 

[airfathaaaaa]

2014 and 2015 AMD games did not sell well.

Example
Battlefield front, Star Wars, Dirt Rally, Hitman 2016.

Compare with Nvidia

MGS V, Witcher 3, Fallout 4, Dying light, Mad Max and Rise of the Tomb Raider. They sold well because they are pro Nvidia game.

I wont buy any game which is totally pro AMD game.

The Divison vs hitman 2016

Division already had 1 million user playing beta from Steam which do not include uplay and other 3rd party.

Hitman 2016 do not even have proper 50K user that played beta.

hmm i dont know
13million copies bf along with 12million on sw copies 140.000 copies for a rally game.. and hitman which just started selling(i mean lol)
and then you comes and argue that mgs v with 5.000.000 copies witcher 3 with 6m copies fallout 4 12m copies dying light with 3.2m copies mad max with 1m copies and rotr that has just passed 1m copies
sold better? sure if you sum it up they did sold better but on raw pur mathematics you are way off

as for the division well numbers never lie
http://steamspy.com/app/414460
check the audience total yet another great game destroyed from a pr team that doesnt know what a game is

 

[Silverforce11]

2014 and 2015 AMD games did not sell well.

Example
Battlefield front, Star Wars, Dirt Rally, Hitman 2016.

Compare with Nvidia

MGS V, Witcher 3, Fallout 4, Dying light, Mad Max and Rise of the Tomb Raider. They sold well because they are pro Nvidia game.

I wont buy any game which is totally pro AMD game.

The Divison vs hitman 2016

Division already had 1 million user playing beta from Steam which do not include uplay and other 3rd party.

Hitman 2016 do not even have proper 50K user that played beta.

Mad Max is not GameWorks. Neutral title, performs great on everything.
Rise of Tomb Raider is 2016.

According to Steamspy:

Dirt Rally (270K) vs Project Cars (356K), Project Cars was released way before Dirt Rally.

Hopefully NV comes up with better "Game Ready" drivers for The Division than their current ones, because it's not looking good for Maxwell, and worse for Kepler.

And it doesn't really matter what NV sponsors moving forward, it's going to run great on AMD GCN because these new game engines have to be optimized for GCN on consoles.

Just look at Far Cry 4, NV sponsored, runs faster on GCN. Just Cause 3, NV sponsored, runs great on GCN. Rainbow Six, NV sponsored, runs awesome on GCN.

Heck, even Witcher 3, that blockbuster, runs awesome on GCN when the performance tanking HairWorks is off.

Even the same for Dying Light.

Really, can you believe it's NV sponsored? Hahaha!

And Rise of the Tomb Raider?

Runs great on GCN too.

 

[ShintaiDK]

No, I didn't ask that, you did, my questions are in the post you quoted. So running better on AMD is only because they're crippling nvidia performance right? Using dx12 ASc is trying to cripple nvidia, ok didn't know that.

Who is it again that complains about tessellation? A feature AMD funny enough championed early on.

 

[Shmee]

Guys, back on topic here. And stay civil. Also, no need to talk about banning members, or playing backseat moderator. If there is an issue, use the report post button. Thanks,

Shmee

 

[Silverforce11]

The excuse is already that Fable doesn't use enough async.

Take a wild guess on how the other DX12 titles will end.

It's not an excuse when it comes directly from the developers, Lionhead Studios.

It's a run through of the environment only, no NPCs, ~5% Async Compute.

On an UE4 that's NV sponsored, pretty damn good result for GCN! Look at the R290X destroying the 970.

Add more async compute, and the result will turn out like Ashes. You remember the early Ashes builds? Exactly similar result, Fury X was behind 980TI, while Hawaii wrecked GM204, and Tahiti wrecked Kepler. Beta 2 when they add Async Compute, boom, all GCN jumps up.

 

[DarkKnightDude]

Division already had 1 million user playing beta from Steam which do not include uplay and other 3rd party.

Hitman 2016 do not even have proper 50K user that played beta.

This is a weird argument from you. Hitman is not even fully out till March 8th and you had to pay to get into the beta. You're comparing it with the Division which had a free open beta, plus has been promoted by Ubisoft for three years. The gameplay is also wildly different between the two, one is a third person RPG shooter, and the other is a game about assassinations.

Fallout, MGS 5, Witcher 3 all sold well because they're great games in general. You think if Witcher 3 had AMD features it would have sold differently? Gameplay and story is what sold those games.

Star Wars Battlefront also sold poorly on PC because it was a bad game, nothing to do with its supposed AMD affiliation (which wasn't even advertised in the game). Lack of PC server browser or mods, plus its super casual gameplay turned off a lot of potential buyers. Graphics wise though it ran amazing on all hardware, and looked fantastic.

BF4, AMD promoted, sold millions I might add. However, it sold millions because its a sequel to Battlefield 3. The first Tomb Raider reboot sold 4 million copies on PC and that was heavily AMD promoted. You think those 4 million copies were sold because they want to see AMD features or something? They want to play the game.

Anyways I'm getting off topic here, and please come up with a better argument.

 

[Pinstripe]

It might also simply have to do with the fact that SW: Battlefront and Hitman 2016 are shitty games.

All these DX12 showcase "games" now coming out over the next couple of months are crap, barring Quantum Break, but that's a Windows Store exclusive, ergo a big no no.

 

[Leadbox]

Who is it again that complains about tessellation? A feature AMD funny enough championed early on.

If we're going to apply 64 factors of it to flat surfaces just because, then someone is bound to complain. Who is it that claimed full dx12 support and more 12_1 but stumbles at the first hurdle?

 

[ShintaiDK]

It's not an excuse when it comes directly from the developers, Lionhead Studios.

It's a run through of the environment only, no NPCs, ~5% Async Compute.

On an UE4 that's NV sponsored, pretty damn good result for GCN! Look at the R290X destroying the 970.

Add more async compute, and the result will turn out like Ashes. You remember the early Ashes builds? Exactly similar result, Fury X was behind 980TI, while Hawaii wrecked GM204, and Tahiti wrecked Kepler. Beta 2 when they add Async Compute, boom, all GCN jumps up.

So the excuse is not enough async. How many times do you think that will be used as an excuse in future DX12 titles? The problem is you take AOTS as the standard for DX12. Yet its very biased towards one vendor and from the looks of it, it will be nothing but a benchmark in its entire life. The exact same was done with DX11 back in the days. And it backfired epically. Just as the FX dreams about now everything would change to support this minority group.

 

[ShintaiDK]

If we're going to apply 64 factors of it to flat surfaces just because, then someone is bound to complain. Who is it that claimed full dx12 support and more 12_1 but stumbles at the first hurdle?

I think you need to read up on DX12 specification

 

[Silverforce11]

So the excuse is not enough async. How many times do you think that will be used as an excuse in future DX12 titles? The problem is you take AOTS as the standard for DX12. Yet its very biased towards one vendor. The exact same was done with DX11 back in the days. And it backfired epically. Just as the FX dreams about now everything would change to support this minority group.

Every single DX12 game announced has mentioned using DX12 Async Compute.

Look up Quantum Break and Gears of Wars, do a search of those and Async Compute. Google will find you the article where their developers talk about using that feature to accelerate graphics effects.

Heck, even Rise of the Tomb Raider, the DX11 port, their devs said their DX12 build uses Async Compute as well as their Xbone build. It was disabled for the port PC (because $$ GameWorks), but they may enable it via a patch.

There's also Mirror's Edge, Deus Ex. Not exactly indie titles.

Maybe the only one that has not, is Kings of Wushu, the NV sponsored title.

 

[ShintaiDK]

Yet you dont know how these perform yet. We got 2 benchmarks so far. fable and AOTS. Both showing 2 completely different outcomes. Yet you already decided. See the problem?

You, like others, have betted the house on AOTS being the performance standard.

 

[Silverforce11]

Yet you dont know how these perform yet. We got 2 benchmarks so far. fable and AOTS. Both showing 2 completely different outcomes. Yet you already decided. See the problem?

They don't show completely different outcomes. You are only looking at Fury X vs 980Ti and ignore everything else.

How do AMD normally perform in UE4 games, do you know? Now compare that to UE4 powered Fable DX12. Mind, blown.

Here's why I've already decided. Mantle ---> Vulkan/DX12. Multi-Thread Rendering and Multi-Engine Queues. Made for GCN to shine.

Doesn't take a genius to link it, but you were quite vocal about that impossibility even back in 2013 when Mantle was released.

 

[Leadbox]

The excuse is already that Fable doesn't use enough async.

Take a wild guess on how the other DX12 titles will end.

Are you one those who only look at 980Ti and infers all else from there? Remove the Ti from that chart as there's a lot more of all the other nvidia cards out there in use vs the Ti, do you still see the picture you're trying to paint after that?

 

[ShintaiDK]

They don't show completely different outcomes. You are only looking at Fury X vs 980Ti and ignore everything else.

How do AMD normally perform in UE4 games, do you know? Now compare that to UE4 powered Fable DX12. Mind, blown.

Here's why I've already decided. Mantle ---> Vulkan/DX12. Multi-Thread Rendering and Multi-Engine Queues. Made for GCN to shine.

Doesn't take a genius to link it, but you were quite vocal about that impossibility even back in 2013 when Mantle was released.

It look over 2 years before Mantle ended up transformed into Vulcan. Mantle itself is dead. And DX12 is a different beats.

But what's different than with DX10/11 where the outcome turned out entirely different? All your data depends on 1 single benchmark, while you disapprove the other benchmark already as "Nvidia favoured".

AOTS sold 20K copies so far. That's what you base the success on.

 

[sontin]

If we're going to apply 64 factors of it to flat surfaces just because, then someone is bound to complain. Who is it that claimed full dx12 support and more 12_1 but stumbles at the first hurdle?

AMD helped to develop the Heaven benchmark and promoted it in 2009. Look who "stumbles" at the first hurdle back then.

And what is wrong with "64 factors of it to flat surfaces just because" and "lets compute nonsense like missle for every unit"?!

/edit: This "game" doesnt look better than Starcraft 2 from 2009(?).