Gameworks, Mantle and a pot calling a kettle black

[LtMatt]

Star swarm is not consistent enough so in all honesty that means nothing.

 

[_Alatar_]

Both are from the same custom scenario.

If you do enough testing you'll see that runs using that custom scenario on the same system are within a couple of frames of each other.

 

[LtMatt]

Both are from the same custom scenario.

If you do enough testing you'll see that runs using that custom scenario on the same system are within a couple of frames of each other.

Using that custom scenario takes out the motion blur part of the bench which is the most demanding part as it renders the scene 9 times when that occurs making the most of Mantles draw call advantage. That happens in the follow mode but not in the custom csv i just used.

 

[_Alatar_]

No it does not take the motion blur out.

It's still there every bit as much as it is in follow. You just can't see it as clearly because the camera isn't doing extremely fast movements. But it's still there, it does all the work it does in follow mode, but due to the static camera you can't see the benefit as much.

That custom is more demanding on the CPU than follow is.

 

[LtMatt]

Lol. Yes it does. You can't see it because it's not there. Looks like we're going to have to agree to disagree, again. I'm out to save the thread derail.

 

[TheRealDeal]

Please don't come back at all.

Never really been away, i was just waiting for a really constructive post like yours to step back into the thread.

 

[GoogalyMoogaly]

I'm not getting drawn back into the GameWorks debate again Greg, I've said my piece on that for now. I can put you in touch with Joel via trust if you want though, you can ask him yourself then.

Well at least Matt outright states that he's not going to get back on topic in the thread. Wonder how happy he'd be if we refused to get back on topic in one of his threads?

Also avoided the question nicely too.

 

[LtMatt]

Well at least Matt outright states that he's not going to get back on topic in the thread. Wonder how happy he'd be if we refused to get back on topic in one of his threads?

Also avoided the question nicely too.

Not getting into another argument no matter how much you or others try to bait, but have fun.

 

[GoogalyMoogaly]

Not getting into another argument no matter how much you or others try to bait, but have fun.

Yeah, people often bow out of arguments when they don't have an answer.

So Batman: AO performs just as well on AMD as Nvidia?
And Watch Dogs performs just as well on AMD as Nvidia?

And we're saying that GameWorks is sabotaging performance on AMD cards?
They're both GameWorks games right?
So is the problem that AMD users think they should be getting more performance than the respective Nvidia cards and are unhappy because they're getting fairly even performance?

Obviously I don't expect Matt to answer as I think he's made it clear he doesn't have one. But if someone could explain it to me?

 

[triss]

Dont try and introduce logic googly people dont like it

 

[LtMatt]

Yeah, people often bow out of arguments when they don't have an answer.

So Batman: AO performs just as well on AMD as Nvidia?
And Watch Dogs performs just as well on AMD as Nvidia?

And we're saying that GameWorks is sabotaging performance on AMD cards?
They're both GameWorks games right?
So is the problem that AMD users think they should be getting more performance than the respective Nvidia cards and are unhappy because they're getting fairly even performance?

Obviously I don't expect Matt to answer as I think he's made it clear he doesn't have one. But if someone could explain it to me?

AO performs just as well at max settings, as x8 AA favours AMD architecture. However using FXAA budget range Nvidia cards beat mid tier AMD cards and mid range Nvidia cards beat high end AMD cards. That's the difference. Doesn't affect the high end, but does affect the low to mid end card users.

I've also said in this very thread that i don't believe Nvidia designed GameWorks to gimp AMD performance so really there is no need to keep baiting me.

 

[GoogalyMoogaly]

AO performs just as well at max settings, as x8 AA favours AMD architecture. However using FXAA budget range Nvidia cards beat mid tier AMD cards and mid range Nvidia cards beat high end AMD cards. That's the difference. Doesn't affect the high end, but does affect the low to mid end card users.

I've also said in this very thread that i don't believe Nvidia designed GameWorks to gimp AMD performance so really there is no need to keep baiting me.

Weren't you out?

Also didn't you start a whole thread about GameWorks intentionally gimping AMD performance?

 

[LambChop]

Loving this thread so far.

 

[Martini1991]

AO performs just as well at max settings, as x8 AA favours AMD architecture. However using FXAA budget range Nvidia cards beat mid tier AMD cards and mid range Nvidia cards beat high end AMD cards. That's the difference. Doesn't affect the high end, but does affect the low to mid end card users.

I've also said in this very thread that i don't believe Nvidia designed GameWorks to gimp AMD performance so really there is no need to keep baiting me.

How do we know it's not Nvidia cards running much faster than they should be, rather than the AMD cards running slow?

Hence Nvidia's dramatic MSAA drop in performance (Yes, I know, AMD loses less, but we're talking a much higher drop than is the norm for Nvidia.)

 

[LtMatt]

Weren't you out?

Also didn't you start a whole thread about GameWorks intentionally gimping AMD performance?

I was but didn't want you thinking i wouldn't answer, and yes i posted several news articles examining GameWorks.

How do we know it's not Nvidia cards running much faster than they should be, rather than the AMD cards running slow?

Hence Nvidia's dramatic MSAA drop in performance (Yes, I know, AMD loses less, but we're talking a much higher drop than is the norm for Nvidia.)

AMD's archicture handles AA better than Kepler. That shows in more games than just batman. As to why that specifically is i have no idea.

When AA is enabled do you see a 7970 beating a titan or a 780TI? Or a 7770 beating a 760?

 

[_Alatar_]

Lol. Yes it does. You can't see it because it's not there. Looks like we're going to have to agree to disagree, again. I'm out to save the thread derail.

I'm going to quote myself from the other thread here because I don't want people thinking that matt actually knew what he was talking about and I don't want to type everything again:

lmao

absolutely clueless.

Let me tell you a secret about game development. Assuming an effect or a mesh is on the screen being rendered it doesn't matter how well the player sees it. That exact same effect/mesh still uses the exact same draw call/CPU time budget because you as a developer have told it to be rendered in the scene.

Imagine I want a million 2d butterly meshes on my screen at the same time. All moving constantly. Now that I have those added I have a crapton of draw calls to deal with. In my engine editor I can set the player camera very close to the butterlies or very far away from them. Being closer to the meshes the player can see them better. However this does not change the amount of draw calls that have to be handled.

Just because the gamer can't see X as well as before due to distance, camera movement etc. does not mean that it's not being rendered exactly the same way as it was before. (usually this causes GPU time issues which is why most of the time you want different LOD meshes for your models)

If you wanted to remove a ton of draw calls from this star swarm demo you'd have to change the motion blur implementation (which doesn't change unless you change settings, motion blur implementation is the exact same regardless of scenario, otherwise you'd see the RTS scenario being the easiest to run and follow being the hardest). Or alternatively you could group a ton of individual meshes together into single big meshes, which obviously isn't happening either.

Matt you're just plain wrong. Your rudimentary understanding of game engines rendering something is completely wrong.

Let's prove that.

Follow [Extreme settings]

Spoiler

74fps

Follow [Extreme settings without motion blur]

Spoiler

120fps

Custom Scenario [Extreme settings]

Spoiler

51fps

Custom Scenario [Extreme settings without motion blur]

Spoiler

80fps

Follow blur to no blur fps increase: 62%

Custom blur to no blur fps increase: 57%

It's the exact same fps hit between the modes. The same motion blur is being rendered every time. And as you can see the custom scenario is actually the more stressful one out of the two. The small variation in percentages can be attributed to the inconsistency of the follow preset.

TL;DR same motion blur is being rendered every time, there's no 5 minute drop off when using the CPU overhead improvements in NV's DX11 drivers and Star swarm is now slower on AMD mantle than Nvidia DX11.

End of story for that tangent.

 

[Martini1991]

AMD's archicture handles AA better than Kepler. That shows in more games than just batman. As to why that specifically is i have no idea.

When AA is enabled do you see a 7970 beating a titan or a 780TI? Or a 7770 beating a 760?

Yeah, that doesn't actually address or answer my question.

 

[LtMatt]

Yeah, that doesn't actually address or answer my question.

I don't really know how to answer your question. All i can do is explain why AMD are able to catch up when X8 AA is applied due to fact GCN handles AA better than Kepler. I still don't understand how a 770 is able to best a 290X though when AA is not applied. Can you explain that?

 

[Gregster]

AO performs just as well at max settings, as x8 AA favours AMD architecture. However using FXAA budget range Nvidia cards beat mid tier AMD cards and mid range Nvidia cards beat high end AMD cards. That's the difference. Doesn't affect the high end, but does affect the low to mid end card users.

I've also said in this very thread that i don't believe Nvidia designed GameWorks to gimp AMD performance so really there is no need to keep baiting me.

It was you who started numerous threads on GameWorks and how it gimps performance on AMD cards but as has been shown and proven in our own OcUK Batman Arkham Origins bench thread that this isn't the case. Batman was the original talking point and I believe Batman was released in April and extremetech did the first article some time after. Now clearly if GameWorks (what the whole argument is about) is gimping performance on AMD (intentional or not), At max settings, nVidia cards would still have the same massive advantage, purely based on GameWorks being the accused.

So why does it do it on lower settings but not at max settings? Surely GameWorks would cripple AMD cards at max details as well, as I doubt nVidia can pick and choose what settings to cripple AMD cards.

Edit:

As for baiting, I have seen so many silly gifs/accusations of being a fanboy and starting thread after thread about how GameWorks is bad, I will go with the title of this thread and say "pot, kettle, black".

 

[LtMatt]

It was you who started numerous threads on GameWorks and how it gimps performance on AMD cards but as has been shown and proven in our own OcUK Batman Arkham Origins bench thread that this isn't the case. Batman was the original talking point and I believe Batman was released in April and extremetech did the first article some time after. Now clearly if GameWorks (what the whole argument is about) is gimping performance on AMD (intentional or not), At max settings, nVidia cards would still have the same massive advantage, purely based on GameWorks being the accused.

So why does it do it on lower settings but not at max settings? Surely GameWorks would cripple AMD cards at max details as well, as I doubt nVidia can pick and choose what settings to cripple AMD cards.

Yes and i said i changed my mind somewhat on what GameWorks intention was. Not sure what else you want from me really.

As I've already explained at x8 AA AMD cards are able to overpower any advantage that is present when only FXAA only is used, for whatever reason. Under no circumstance is it normal for a 770 to be beating a 290X though, i think we can all agree that.