CPR W3 on Hairworks(Nvidia Game Works)-'the code of this feature cannot be optimized for AMD'

[Aurhinius]

Well, that is my take on it and as for this church mentality, I noticed a post from you earlier accusing people of always playing this church mentality card but the only person I have ever seen put GPU's and the Church together is you lol. I think you are reading your own posts and getting wound up bud.

To try and answer your question about gimping. I understand that the 960 is very good at tessellation (pretty much like the 285 is better for AMD) and deals with tessellation far better than Kepler, so with The Witcher 3 using big amounts of Tess, I imagine this is why it is looking better on some of the graphs we have seen. I don't for one minute think that nVidia are gimping Kepler cards and just the fact that they don't deal with Tess so well.

@ Tommy

Like I said to Humbug, that is the way I see it and enjoy your new nVidia card/s

Wasnt the 700 series marketing touting how much better kepler tessellation was than the 290 series yet the 290s are beating 780s in benchmarks i hear. Perhaps nvidia needs to introduce a tessellation slider like amd.
http://www.geeks3d.com/20131108/geforce-gtx-780-ti-the-king-of-tessellation/

Hmmm.... according to that site the 780s should be storming over the 290s but we arent seeing that are we?

Terchreporthttp://techreport.com/review/25611/nvidia-geforce-gtx-780-ti-graphics-card-reviewed/4

So what is up with the 780s tessellation performance?

 

[Silent_Scone]

Sorry, was under the impression the Nixxes example was easy enough to understand.

@greg, thanks, I am.

Oooh, nice one.

It was easy to understand considering all you were doing is relaying what you read happened. It's also the only example to date, because nobody has the kind of resources to coactively work like that unless there is money behind it.

 

[Gregster]

Wasnt the 700 series marketing touting how much better kepler tessellation was than the 290 series yet the 290s are beating 780s in benchmarks i hear. Perhaps nvidia needs to introduce a tessellation slider like amd.
http://www.geeks3d.com/20131108/geforce-gtx-780-ti-the-king-of-tessellation/

Hmmm.... according to that site the 780s should be storming over the 290s but we arent seeing that are we?

I followed Kepler with eagle eyes, as I treated myself to a pair of Titans and don't recall tessellation being a selling point. Not saying it wasn't mentioned but I honestly don't recall that. There is tessmark, which shows how well each card does in tessellation and Kepler is quite down the list compared to Maxwell.

 

[LambChop]

Maybe someone with a 780 runn tessmark with an old driver then run with a new one and see what the differences are ?

 

[Silent_Scone]

http://www.forbes.com/sites/jasonev...ng-about-the-witcher-3-and-nvidias-hairworks/

 

[Aurhinius]

I followed Kepler with eagle eyes, as I treated myself to a pair of Titans and don't recall tessellation being a selling point. Not saying it wasn't mentioned but I honestly don't recall that. There is tessmark, which shows how well each card does in tessellation and Kepler is quite down the list compared to Maxwell.

0kay greg here is tessmarkhttp://techreport.com/review/27702/nvidia-geforce-gtx-960-graphics-card-reviewed/4

Check the 960 score then the 770 score. The 770 beats the 960 in tessellation. Now lets look back at the fps charts for witcher 3.
http://www.guru3d.com/articles_pages/the_witcher_3_graphics_performance_review,5.html

So 770 has better tessellation but performs 7fps slower than the 960 at 1600x1200. You cant tell me that doesnt look right!

Checking tessmark numbers the 780ti scores higher than a 970 in tessellation. Yet in the benchmarks scores 9fps less......

Something is not right i suspect either in w3 tessellation implementation or in drivers for the 7 series cards.

 

[r7slayer]

Wait so if nVidia provide source code to developers who request it.

'GameWorks source code is available to licensees. We provide source code, under license, to developers who request it

why dont AMD make a small team under the impression they are some other company, then request to use game works under license and actually get a game made and use game works and optimize for AMD cards while getting so see the source code at the same time?

BOOM i'm a genius! lol

 

[r7slayer]

http://www.forbes.com/sites/jasonev...ng-about-the-witcher-3-and-nvidias-hairworks/

he uses GTA 5 as an example where both technologies from both camps was used in this game and optimised quite well for both camps but what he fails to mention is that this is pretty much the only game where both camps have worked closely with the developer and implemented their tech and optimized.

Only tech in GTA 5 from nVidia was their multisampling and shadows from my knowledge anyways which AMD can't use the multisample so don't know if that would cripple AMD hardware and nVidia's shadows looked really bad and performed worse that rockstar's own shadowing methods. AMD's shadows look slightly better than nVidia's but still took more of a performance hit than rockstar's version. So i don't think we should use this game as an example.

Think AMD are more about where nVidia jump on devs to work close with them offer them contracts and shut AMD out. This didn't happen with GTA 5 obviously...

 

[Gregster]

0kay greg here is tessmarkhttp://techreport.com/review/27702/nvidia-geforce-gtx-960-graphics-card-reviewed/4

Check the 960 score then the 770 score. The 770 beats the 960 in tessellation. Now lets look back at the fps charts for witcher 3.
http://www.guru3d.com/articles_pages/the_witcher_3_graphics_performance_review,5.html

So 770 has better tessellation but performs 7fps slower than the 960 at 1600x1200. You cant tell me that doesnt look right!

Checking tessmark numbers the 780ti scores higher than a 970 in tessellation. Yet in the benchmarks scores 9fps less......

Something is not right i suspect either in w3 tessellation implementation or in drivers for the 7 series cards.

No idea lol But nVidia are investigating it. Maybe they just haven't spent that long on Kepler/Fermi as they have on Maxwell optimisations... I honestly have no idea but they wouldn't purposefully cripple performance.

 

[Mei]

Wasnt the 700 series marketing touting how much better kepler tessellation was than the 290 series yet the 290s are beating 780s in benchmarks i hear. Perhaps nvidia needs to introduce a tessellation slider like amd.
http://www.geeks3d.com/20131108/geforce-gtx-780-ti-the-king-of-tessellation/

Hmmm.... according to that site the 780s should be storming over the 290s but we arent seeing that are we?

Terchreporthttp://techreport.com/review/25611/nvidia-geforce-gtx-780-ti-graphics-card-reviewed/4

So what is up with the 780s tessellation performance?

i dont think that slider would be needed if these games were optimised prior to going out, they have lots of different hardware to test on its not hard to get, they know how the game is going to perform, yet they still embed crazy high settings in these games that have little to no visual improvement

fast patches are great but it is still ruining the experience for gamers, people who have sometimes waited a long time for these things to come out, why should they have to wait until half way through the game before they get a nice experience, its not like a lot of these games have high replay value

i duno i dont think its good enough at all
/my rant for the day

 

[LambChop]

http://www.forbes.com/sites/jasonev...ng-about-the-witcher-3-and-nvidias-hairworks/

lol at AMD and Huddy especially, my word.

 

[humbug]

No idea lol But nVidia are investigating it. Maybe they just haven't spent that long on Kepler/Fermi as they have on Maxwell optimisations... I honestly have no idea but they wouldn't purposefully cripple performance.

They do cripple thier own performance by using excessive amounts of Tessellation, Nvidia users could do with their own Tessellation slider.

As for Diminishing Kepler performance, if Kepler falls further and further behind Maxwell it will ramp up Maxwell sales, especially all those users seeing a lower mid range card like the 960 matching and over taking their 780.
Nvidia users aren't going to complain, they will sit there for an eternity telling themselves its all fine, Nvidia will fix it, Nvidia will look after me. A 780 is twice the card, Greg.

 

[LambChop]

They do cripple thier own performance by using excessive amounts of Tessellation, Nvidia users could do with their own Tessellation slider.

As for Diminishing Kepler performance, if Kepler falls further and further behind Maxwell it will ramp up Maxwell sales, Nvidia users aren't going to complain, they can't do no wrong.

Nvidia users dont need a slider, dont mix up excellent tessellation performing cards with rubbish ones.

You keep banging the same drum re: nvidia users/nvidia themselves - It's becoming a bit tedious.

 

[Silent_Scone]

Humbug, the developers are able to implement the tessellation however they want, it's not like they input the link libraries and BAM Geralt gets a swishy head of hair. LOL. "Huddy is saying well why are they using line tessellation for it's incredibly inefficient."

Ok that's nice why are you telling us this, take it up with CDRed.

 

[humbug]

The Tessellation is backed into Hairworks. Developers have no control of it. ^^^^

Nvidia users dont need a slider, dont mix up excellent tessellation performing cards with rubbish ones.

You keep banging the same drum re: nvidia users/nvidia themselves - It's becoming a bit tedious.

Over tessellation will reduce the performance on any card, no matter how much throughput it has.
AMD's performance actually end's up faster than Nvidia equivalent's because AMD's drives can override over tessellation.

 

[Aurhinius]

No idea lol But nVidia are investigating it. Maybe they just haven't spent that long on Kepler/Fermi as they have on Maxwell optimisations... I honestly have no idea but they wouldn't purposefully cripple performance.

Agree totally but it is something they shouldnt be allowed to ignore because they have a large customer base with 7 series cards. I would be pretty annoyed if i had bought a titan or titan black right now as they are performing worse than a 780ti and close to a 960

 

[LambChop]

The Tessellation is backed into Hairworks. Developers have no control of it. ^^^^



Over tessellation will reduce the performance on any card, no matter how much throughput it has.
AMD's performance actually end's up faster than Nvidia equivalent's because AMD's drives can override over tessellation.

You are overriding something that was not intended. AMD only have a slider due to their tessellation not being up to par. Nvidia does not need a slider to reduce it if its over-tessellated. It can easily cope with the levels that is applied.

 

[Aurhinius]

Nvidia users dont need a slider, dont mix up excellent tessellation performing cards with rubbish ones.

You keep banging the same drum re: nvidia users/nvidia themselves - It's becoming a bit tedious.

Sorry but did you check tessmark scores I posted? A 970 scores less than a 780ti in tessmark but in actual game performance the 970 outperforms the 780ti by 9 fps. Approx 20% Better.


I dont think nvidia are doing it on purpose but it might be a case of neglect. Who knows but we should beat the drum so that it doesnt get ignored and nvidia acknowledge and fix their own issues with 7 series performance.
if they cant fix it i would like an explanation other than a 9 series card is newer thus better

 

[Mei]

You are overriding something that was not intended. AMD only have a slider due to their tessellation not being up to par. Nvidia does not need a slider to reduce it if its over-tessellated. It can easily cope with the levels that is applied.

lol you must have missed the complaints from people with £900 titan-x's saying hairworks was killing their frames until the recent patch

 

[Ayahuasca]

I was playing with everything maxed before the patch and hairworks wasn't killing my frames at 1440p.