Does Gameworks kill performance? Lets find out...as a community

[Howling]

lol at all this "not relevant" or "of course performance goes down" **** from nVidia defenders.
AA makes performance go down, but it's fairly well documented the effect on framerates that different levels of AA have using the different methods of AA. GameWorks makes performance down, but it's not documented to what extent - the OP is clearly just trying to get an understanding of how much performance hit these effects have. It's a setting, like any other - just because it says "nVidia" at the front doesn't mean it needs defending. Aside anything else I'm not sure what you're defending it from?

 

[D.P.]

lol at all this "not relevant" or "of course performance goes down" **** from nVidia defenders.
AA makes performance go down, but it's fairly well documented the effect on framerates that different levels of AA have using the different methods of AA. GameWorks makes performance down, but it's not documented to what extent - the OP is clearly just trying to get an understanding of how much performance hit these effects have. It's a setting, like any other - just because it says "nVidia" at the front doesn't mean it needs defending. Aside anything else I'm not sure what you're defending it from?

The performance impact will entirely depend on what effect is used, what the rest of the scene entails, drivers, hardware, and what the developer configured.


There is nothing insightful you can really get and no general summary unlike something like AA which is completely different.

Just because people think this is waste of time doesn't mean they are defining anything. you re the one trying to defend a fruitless waste of time.

 

[Howling]

Ok

 

[Mei]

lol DP who are you to say what is a waste of time
i think the results could be quite interesting
id like to see the fps cost on different games & diff systems
its all good!?

 

[D.P.]

lol DP who are you to say what is a waste of time
i think the results could be quite interesting
id like to see the fps cost on different games & diff systems
its all good!?

It wont give you any insight in to future games for starters.
And it tells you absolutely nothing about the performance optimization of GamesWorks.

The title of the thread is "Does Gameworks kill performance? Lets find out...as a community". The answer to that is no, thread over really.


EDIT:
Personal I would love to see a rigorous comparison of GameWorks features like hair-works pitted against alternatives like TressFX and a developers home-brew solution to see visual and performance trade offs. The only informaiton I know of is form some developers on Beyond3D who said they were happy with GW performance bit no specifics.

 

[Mei]

wow you are not judge and jury of the world!!

i dont see why people need to tiptoe so much just incase they upset a fan of either camp
if the thread is so worthless you dont need to read it do you!?

 

[Rroff]

lol at all this "not relevant" or "of course performance goes down" **** from nVidia defenders.
AA makes performance go down, but it's fairly well documented the effect on framerates that different levels of AA have using the different methods of AA. GameWorks makes performance down, but it's not documented to what extent - the OP is clearly just trying to get an understanding of how much performance hit these effects have. It's a setting, like any other - just because it says "nVidia" at the front doesn't mean it needs defending. Aside anything else I'm not sure what you're defending it from?

You kind of answer it in your own post - you can get a meaningful metric for measuring something like the performance hit from AA but just turning GameWorks features on/off as a comparison doesn't give you a meaningful metric as to whether the effects themselves are worth the performance hit or not or whether they are efficiently implemented or not.

Its kind of like taking Doom 3 - turning off the lighting system and then saying yes real time stencil shadows kill performance... well no ****.

 

[Nexus18]

Precisely, if you don't like the thread, don't read it then? Simples...

 

[D.P.]

wow you are not judge and jury of the world!!

i dont see why people need to tiptoe so much just incase they upset a fan of either camp
if the thread is so worthless you dont need to read it do you!?

I'm not being a judge, i'm purely pointing out that it is a needless waste of time that wont illuminate anything.

Rroff gave a good summary why.

Turning off effects and seeing the FPS go up doesn't tell you anything interesting or anything you could apply to future games, or other games.


Of course if you own a game it is useful to change settings to see what performance and visual changes there are. But why limit that to GamesWorks. why not test games at all kind of different detail setting. What does shadow detail change, what about reflection quality, draw distance, texture quality?
Why this fascination with testing GameWorks effects only?


There are really 2 separate questions here:
1) Is GamesWorks reasonably optimal compared to other middleware libraries and developers own solutions? Is there a strong vendor bias?
2) What performance and visual changes by switching quality of different game effects, GamesWorks or otherwise.



This thread is not doing anything useful like that. The thread title says it all, and its only purpose is another pointless nvidia defamation.

Switching on and off GamesWorks effects is pointless, unless you want to make some kind of statement about your brand affection.

For the record, I would love to see a comprehensive comparison of GameWorks., my question 1 above. No one on OCUk can answer that though.

 

[Howling]

I'm not being a judge, i'm purely pointing out that it is a needless waste of time that wont illuminate anything.

Rroff gave a good summary why.

Turning off effects and seeing the FPS go up doesn't tell you anything interesting or anything you could apply to future games, or other games.


Of course if you own a game it is useful to change settings to see what performance and visual changes there are. But why limit that to GamesWorks. why not test games at all kind of different detail setting. What does shadow detail change, what about reflection quality, draw distance, texture quality?
Why this fascination with testing GameWorks effects only?

I think everyone knows roughly what to expect when they turn Shadows to Ultra and Reflection to Ultra in terms of performance, it's usually a large hit for little improvement over high, but putting them on High yields a reasonable visual improvement over Med for a smaller impact. These are settings we are used to seeing and know the first settings to turn down if we're having performance issues. GameWorks is an unknown "new" setting to play with, seeing what effects do and their performance cost is always helpful.

There are really 2 separate questions here:
1) Is GamesWorks reasonably optimal compared to other middleware libraries and developers own solutions? Is there a strong vendor bias?
2) What performance and visual changes by switching quality of different game effects, GamesWorks or otherwise.



This thread is not doing anything useful like that. The thread title says it all, and its only purpose is another pointless nvidia defamation.

Switching on and off GamesWorks effects is pointless, unless you want to make some kind of statement about your brand affection.

For the record, I would love to see a comprehensive comparison of GameWorks., my question 1 above. No one on OCUk can answer that though.

I disagree, I think it's working towards answering question 2 - although question 2 is a multipart question itself. But yes, those 2 questions would be nice to know the answers to ;D

 

[Avenged7Fold]

I want to see the IQ before and after as well as the relevant frame rates so i know what I can get out of what hardware. It has been mentioned that gameworks obviously has a significant performance hit, it is worth seeing what you get for that performance hit.

It doesn't tell you what's at fault or what other games bring but as someone who is considering the purchase of FO4 as well as considering possible upgrades, this is useful for me and likely many others looking to play FO4.

If the performance is poor regardless off effects and IQ is mediocre, then i cant see myself purchasing this game regardless of game-play, cause y'know, why buy a game which you cant get motivated to play due to low frame rates?

I think everyone knows roughly what to expect when they turn Shadows to Ultra and Reflection to Ultra in terms of performance, it's usually a large hit for little improvement over high, but putting them on High yields a reasonable visual improvement over Med for a smaller impact. These are settings we are used to seeing and know the first settings to turn down if we're having performance issues. GameWorks is an unknown "new" setting to play with, seeing what effects do and their performance cost is always helpful.

Precisely.

Spend money and time upgrading and overclocking, it makes sense to want to know what you get out of your hardware and what you get back for the performance hit. Like i said:

Problem with a thread like this is that some of the people are set on a witch hunt and others are out to ardently defend, where as it should be looked at subjectively. eg. This is the performance drop, this is what it looks like with it on and with it off.

Personal judgements on why it is can be reserved for after that and in reality, reasons for the performance hit matter little compared to what the performance hit is and what you get for it.

 

[D.P.]

I think everyone knows roughly what to expect when they turn Shadows to Ultra and Reflection to Ultra in terms of performance, it's usually a large hit for little improvement over high, but putting them on High yields a reasonable visual improvement over Med for a smaller impact. These are settings we are used to seeing and know the first settings to turn down if we're having performance issues. GameWorks is an unknown "new" setting to play with, seeing what effects do and their performance cost is always helpful.



I disagree, I think it's working towards answering question 2 - although question 2 is a multipart question itself. But yes, those 2 questions would be nice to know the answers to ;D

If the thread is aiming to answer question 2 then the title and all posts referring to Games Works should be ignored and everyone can test different effects and quality settings, including GW effects.

Personally I would love for GamesWorks to be shown to be very poorly optimized because it will force Nvidia to put a lot more resources into its development and it will help everyone, but I don't think that is the case at all. Nvidia have a large team of experts, countless doctorates in physics, maths, CS, and have hired some of the best developers, picking up people from crytek, Valve etc. Nvidia pay well and try to buy in the best, they have the financial resources to ensure things are done right. that is a recipe for generating good software, no guarantee of course, bad management or corporate atmosphere can have a big impact. The only things I have heard form developers is that performance is good and GW does what it says on the tin.

 

[Rroff]

If the thread is aiming to answer question 2 then the title and all posts referring to Games Works should be ignored and everyone can test different effects and quality settings, including GW effects.

Personally I would love for GamesWorks to be shown to be very poorly optimized because it will force Nvidia to put a lot more resources into its development and it will help everyone, but I don't think that is the case at all. Nvidia have a large team of experts, countless doctorates in physics, maths, CS, and have hired some of the best developers, picking up people from crytek, Valve etc. Nvidia pay well and try to buy in the best, they have the financial resources to ensure things are done right. that is a recipe for generating good software, no guarantee of course, bad management or corporate atmosphere can have a big impact. The only things I have heard form developers is that performance is good and GW does what it says on the tin.

nVidia also maintain a significant resource of game programming material i.e. https://developer.nvidia.com/gpugems/GPUGems/gpugems_pref01.html so it would look pretty bad if their own implementations were second rate - though I'm still not convinced they are as optimal as they could be.

 

[fs123]

....
This thread is not doing anything useful like that. The thread title says it all, and its only purpose is another pointless nvidia defamation.

Switching on and off GamesWorks effects is pointless, unless you want to make some kind of statement about your brand affection.

For the record, I would love to see a comprehensive comparison of GameWorks., my question 1 above. No one on OCUk can answer that though.

In case you hadn't noticed, the op owns a 980ti and is a fan of Nvidia so your claim of defamation is baseless. Unless you work for Nvidia or are a blind fan you shouldn't have a problem with a thread that aims to find out how much performance is lost by enabling Gameworks features.

The thread is valid since Gameworks is not something where developers create code differently for different games. Any game which uses these libraries should exhibit the same or similar level of performance degradation per feature so it's worth investigating.

 

[D.P.]

In case you hadn't noticed, the op owns a 980ti and is a fan of Nvidia so your claim of defamation is baseless. Unless you work for Nvidia or are a blind fan you shouldn't have a problem with a thread that aims to find out how much performance is lost by enabling Gameworks features.

I don;t care what the OP owns. I'm just pointing out the obvious flaws with the proposed testing.

The thread is valid since Gameworks is not something where developers create code differently for different games. Any game which uses these libraries should exhibit the same or similar level of performance degradation per feature so it's worth investigating.

That is where you are completely wrong because the performance is highly related to the way the develops implement the features, configure the options and other resources the GPU has to expend rendering the rest of the scene. You cannot compare GW performance between games, there is no predicative value there. Who know what kind of quality options the developer chooses, who knows what else the game engine is rendering. For example, if the developer heavily uses tessellation for soemthign like terrain rendering and then turns up tessellation quality high on hair works then the effect will be highly dependent on the GPUS's tessellation performance, another developer might not use tessellation at all and turn quality setting down lower with the hairworks effects and get a very different performance outcome, and to further complexity things the game with heavy tessellation requirement might also have extremely heavy fragment shaders that mask any tessellation issues.

you simply cannot compare form one game to another. Some effects will be more computational intensive than others, that is just a fact of life. How that affects game performance is largely down to the developers, and the fact that the effect is from a gamesworks library is irrelevant.


Interactive volumetric fog is a good example of a computationally complex simulation but is highly dependent on a number of developer controlled factors.

 

[Dicehunter]

Didn't take long for this thread to turn toxic

 

[shankly1985]

Didn't take long for this thread to turn toxic

Yep same old same old.

 

[Dygaza]

Game is very cpu heavy

Spoiler

 

[D.P.]

Didn't take long for this thread to turn toxic

Always the same useless rubbish. "Nvidia slept with my wife", "Nvida kills puppies for fun", "GamesWorks gave me cancer". As soon as someone injects some rationality into the mindless bickering people come out and complain?

I would love a thread that compared GamesWorks to alternatives. This thread isn't it.
I don't care if people want to compare the performance hit of different effects, but why limit yourself to GamesWorks unless you have an agenda? and if you don't have an agenda then accept that the suggestion to test all effects is a valid constructive criticism.


People trying to paint the typical "Nvidia is evil" picture is what makes these threads toxic, along with all those that try to support such practices. There is plenty of scope for a constructive and informative thread if people wanted to be objective.

 

[Gregster]

Well the thread was a good idea but thinking on it, AMD can't run the GameWorks effects or at least not all of them but for me it gave a 33% performance hit which didn't bring it under the threshold of unplayable and I like what the effects brings. There is no real mean for testing as there isn't an alternative really (TressFX perhaps?) But if anyone does code a alternative, I am all up for seeing it and comparing.