Digital Foundry Article Technical Discussion Archive [2016 - 2017]

[Recop]

I didn't say no other developer uses it, I just wanted to be clear that not everyone can immediately and for different reasons. We are generalizing words, soI know the point in which you were trying to get at.
tldr;
It's never as simple as just integrating it in just because its available. We've waited a very very long time for Tiled resources to appear in games after it's announcement in 2012.

I certainly trust our resident developer word's on that topic this forum:

#484
#482

Article about execute Indirect:
http://www.dsogaming.com/news/direc...proves-performance-greatly-reduces-cpu-usage/
Our resident moderator Andrew Lauritzen was the one who coded that demo so he can verify the performance gains written about in the article.
I was there in person when Max McCullen presented that demo at Build, so I can verify that the DSO wasn't pulling numbers out of thin air.

And you're making a huge confusion between the XB1 and PC... the XB1 is a console that can be coded to the metal since launch...

So, your points aren't valid here...

 

[Recop]

We can clearly see an advantage of about 56% in previous games, now this advantages has been shortened and is about 15%.

This is a huge boost on XBOX ONE.

This is a nonsense if we consider the respective hardwares... if NXGamer is right, then the only thing that can explain this move are commercial arguments.

 

[Recop]

You missed a few slides especially on the Future work. It's obvious that the XB1 has hardware features that are still not used in today's games. To support these features the console needs a lot of stuff to happen.

The XB1 is not a PC... it's a console...

 

[zed]

We can clearly see an advantage of about 56% in previous games, now this advantages has been shortened and is about 15%.
This is a huge boost on XBOX ONE.

you do realize you're comparing a xb1 vs a xb1s (the xb1s is a faster machine, sure theres no doubt improvements to the engine but the hardwares improved as well)
Also the 56% is due to the 900p vs 720p, i.e. they prolly choose those numbers cause they were common, a more truer resolution for the xbox1 would of been say ~750p but they just rounded it down to the much more common 720p

 

[Rock_Lee]

And you're making a huge confusion between the XB1 and PC... the XB1 is a console that can be coded to the metal since launch...

So, your points aren't valid here...

You are wrong, I don't know whether you have not read the slides or you just are unable to accept any good piece of news concerning xbox one development.

The slides clearly read that those new feaures we are talking about are to be used in the future and on xbox one, In fact the results showed were got without using those features and I can confirm so because I asked Graham whether they had been used in their tests.

 

[iroboto]

And you're making a huge confusion between the XB1 and PC... the XB1 is a console that can be coded to the metal since launch...

So, your points aren't valid here...

This isn't true. Definitely not true. Xbox one has always been using a subset of DX11. Fast semantics was introduced for low level programming but was unavailable for launch. And while it allowed for better control over memory only a handful of features from dx12 were exposed to XBO. (Bundles, much later on async compute, no documentation we've read that they included execute indirect before deprecating Fast Semantics for DX12.X) We have a full thread here on leaked SDK information.

 

[Rock_Lee]

you do realize you're comparing a xb1 vs a xb1s (the xb1s is a faster machine, sure theres no doubt improvements to the engine but the hardwares improved as well)
Also the 56% is due to the 900p vs 720p, i.e. they prolly choose those numbers cause they were common, a more truer resolution for the xbox1 would of been say ~750p but they just rounded it down to the much more common 720p

You are wrong. I'm not comparing xbox one S, the comparison between xbox one s and ps4 is made at the end of the video. The figures given at the beginning are the ones got from xbox one.

As you can see xbox one is labeled as XB1

Xbox one S is labeled as X1S

 

[Recop]

This isn't true. Definitely not true. Xbox one has always been using a subset of DX11. Fast semantics was introduced for low level programming but was unavailable for launch. And while it allowed for better control over memory only a handful of features from dx12 were exposed to XBO. (Bundles, much later on async compute, no documentation we've read that they included execute indirect before deprecating Fast Semantics for DX12.X) We have a full thread here on leaked SDK information.

http://wccftech.com/xbox-one-direct...ute-hardware-specifications-ps4-comparison/5/

 

[Recop]

You are wrong, I don't know whether you have not read the slides or you just are unable to accept any good piece of news concerning xbox one development.

The slides clearly read that those new feaures we are talking about are to be used in the future and on xbox one, In fact the results showed were got without using those features and I can confirm so because I asked Graham whether they had been used in their tests.

Ok, let's say you are right. How can you establish a correlation between this and the resolution upgrade on BF1 ? We can't make a serious discussion on unsupported claims.

Let me do the same thing than you :

"Here we have a clear benchmark that explains why WD2 runs at 1080p on PS4"

See ? I picked a random point in a slide and i just made an unsupported claim... everyone can do that...

Once again, the main part of the slide was how to improve performances by compute. Execute was a detail in their whole presentation.

And the console which is best suited for compute task is the PS4. It's already known since a long time...

 

[djskribbles]

Some things to consider here.
-XB1 had issues with quite a few launch titles running at 720p. BF4 was one of these titles, but +56% is not the typical gap we see on MP games. I believe the general consensus was that the memory was the main culprit for some launch titles.
-All of the comparisons are being done on the XB1S now, which got a slight GPU clock boost over older models.
-Microsoft has marketing rights for BF1, so you can bet that XB1 was the lead console, which has been said by devs on these boards that there is such a thing. Games where MS has marketing rights generally perform slightly better than the gap we typically see on MP games. Yes MS also had marketing rights for BF4, but again, that was a launch title.
-I'm no tech expert, but I believe BF is and always has been CPU limited rather than GPU limited (which may explain why the XB1 version performs slightly better but runs at lower resolution).

Not saying there were no advancements in the XB1, but I think some people are making a bigger deal out of this than there really is.

 

[Recop]

A good API brings a hardware closer to its theoretical peak. And the theoretical peak of the XB1 is still roughly 40% inferior to that of the PS4...

So, i can't understand how people can expect drastric changes from a software update... it's pure logic.

The usual difference we are seeing in MP games is what we can expect from each console.

So, unless people presume that the PS4 API is very bad, the gap will stay the same until the end of the gen...

 

[iroboto]

http://wccftech.com/xbox-one-direct...ute-hardware-specifications-ps4-comparison/5/

Nope. This doesn't prove your point. It may in that XBO gained very little from DX12, but that doesn't necessarily mean it had all those features from launch. Here are some screen caps I took below.

You can see here that fast semantics, the low level variant for XBO was in preview until April 2014, May the following month they recommended that everyone begin using Fast Semantics. Prior to that XBO only had DX11.1, the XBO extensions added were for managing esram.

The remaining features, like Tiled Resources, Execute Indirect and such were added later.
In particular, and to the benefit of your position, executeIndirect (or at least some form of it) was introduced to XBO in May of 2014, titled New MultiDraw APIs -- this would be to replace the existing cruddier DX11 Indirect draws I imagine.
I took a screen cap, note it says similar to multi_draw_indirect on openGL. Unfortunately I do not have the technical knowledge to separate executeIndirect on DX12 from this version on XBO.
But I do see some main differences. Mainly that this version of Multi-draw is to relieve the CPU of batching work over and over again. Where ExecuteIndirect, the GPU is allowed to submit draw calls as well.
And I also know that a great deal of things can change during ExecuteIndirect, where with multiDraw with DX11 you are locked. And there's nothing within this API (at the time) that shows the level of customization that executeIndirect has.

TLDR; i don't believe XBO has had ExecuteIndirect, and didn't until DX12 was released onto XBO. It had multi draw, but I don't think it had ExecuteIndirect.

Continuing,
You notice there that XBO is also just receiving Async Compute, and is now putting Tiled Resources into Preview.

We're pretty positive we saw these features at launch with PS4. KillZone was already running 1 async compute thread as a launch title. I'm also pretty positive they weren't available for DirectX11.1 for PC.

Here is DICE's full slide deck:
http://www.wihlidal.ca/Presentations/GDC_2016_Compute.pdf

In your original question of whether it was there before or not, I believe this somewhat suffices as an answer:

I want to dial back a bit and flip perspectives to your side of the argument. Assume all of this did nothing at all. No performance gain. If the frame rate is tanking, and the resolution has bottomed out since it's dynamic resolution; would you not agree that CPU is likely the culprit? If the CPU is the culprit, there is still up to 30% improvement in optimization here that happened, with a much better looking game over BF4.

Where exactly do you think that 30% came from? You see, we're not interested in proving that PS4 is superior to XBO. We've known that for a very long time. We're interested in the innovation aspect. All you've been tossing up are arguments that fit your needs. Mainly, you are trying to debate PS4 is a better and more powerful system; frankly none of us care.

We're interested in the 30% gain.

 

[iroboto]

Some things to consider here.
-XB1 had issues with quite a few launch titles running at 720p. BF4 was one of these titles, but +56% is not the typical gap we see on MP games. I believe the general consensus was that the memory was the main culprit for some launch titles.
-All of the comparisons are being done on the XB1S now, which got a slight GPU clock boost over older models.
-Microsoft has marketing rights for BF1, so you can bet that XB1 was the lead console, which has been said by devs on these boards that there is such a thing. Games where MS has marketing rights generally perform slightly better than the gap we typically see on MP games. Yes MS also had marketing rights for BF4, but again, that was a launch title.
-I'm no tech expert, but I believe BF is and always has been CPU limited rather than GPU limited (which may explain why the XB1 version performs slightly better but runs at lower resolution).

Not saying there were no advancements in the XB1, but I think some people are making a bigger deal out of this than there really is.

has nothing to do with XB1, just looking at how DICE improved performance for both consoles by 30%. I just put out an idea of how it may have been accomplished. This should not have degenerated to a console war. Cause frankly they're using the same code and running the same architecture.

 

[djskribbles]

has nothing to do with XB1, just looking at how DICE improved performance for both consoles by 30%. I just put out an idea of how it may have been accomplished. This should not have degenerated to a console war. Cause frankly they're using the same code and running the same architecture.

Wasn't including you in my post tbh. Some people are making it seem like the gap is narrowing with superior advancements to XB1, based on the fact that BF1 is a lot closer than BF4. But maybe I'm just getting the wrong impression from their posts.

 

[function]

A good API brings a hardware closer to its theoretical peak.

Errr ... not necessarily.

And the theoretical peak of the XB1 is still roughly 40% inferior to that of the PS4...

Only if you don't know how to calculate percentages.

You don't know why APIs exist, can't calculate percentages, and repeatedly shit the bed desperately attempting to close down conversation.

Why are you here?

 

[mpg1]

Watched that NXGamer vid, here are the numbers:

Average Res (Both Campaign and MP)
Xbox One: 1624x945
PS4: 1780x1000 (+15.8%)

Average FPS (Campaign)
Xbox One: 58.03 (+1.6%)
PS4: 57.1

Average FPS (Multiplayer)
Xbox One: 51.7 (+6.8%)
PS4: 48.4

If you take into account a very slight advantage in average FPS for Xbox One I would say PS4 overall advantage is about 13-15% range.

 

[Recop]

Nope. This doesn't prove your point. It may in that XBO gained very little from DX12, but that doesn't necessarily mean it had all those features from launch. Here are some screen caps I took below.

What's the source of your document ?

Just an example : http://www.hotchips.org/wp-content/...1-epub/HC25.26.121-fixed- XB1 20130826gnn.pdf (2013)

Where exactly do you think that 30% came from? You see, we're not interested in proving that PS4 is superior to XBO. We've known that for a very long time. We're interested in the innovation aspect. All you've been tossing up are arguments that fit your needs. Mainly, you are trying to debate PS4 is a better and more powerful system; frankly none of us care.

Yeah, good question ? Where ? Do you have any idea ?

I'm not interested to prove PS4's superiority... i'm interested by real facts...

Basically, you picked out a random point from a slide and you build a whole argument on it... once again, this is not serious.

 

[Recop]

Only if you don't know how to calculate percentages.

You don't know why APIs exist, can't calculate percentages, and repeatedly shit the bed desperately attempting to close down conversation.

Why are you here?

Yeah, i don't know both. Some links :

- http://www.ign.com/articles/2015/03...e-unlikely-to-boost-the-witcher-3s-resolution

- http://gamingbolt.com/dice-dev-xbox...wont-help-reduce-difference-talks-development

- http://wccftech.com/phil-spencer-directx-1-xbox-one-hardware-graphics/

Also, what was the secret sauce between Uncharted 1 and Uncharted 2 ? Another buzz word or simply a better understanding of the hardware ?

I'm maybe too realistic for you...

 

[Rikimaru]

Xbo has a higher clocked CPU.
Different games have different bottlenecks. Usually disparity is much larger.

 

[helvetica bold]

What kind of performance boot should we see BF1 running on PS4 Pro? Will DICe have the time to really dig into that new hardware?


Sent from my iPad using Tapatalk