Soldato
DX11 result, spec in sig.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Hawx 2 benchmark
- Thread starter maddness
- Start date
Soldato
The even sadder thing is that they could very likely reduce the Tessellation factor for no visible difference and not screw over AMD users as much, but with Nvidia lining their pockets they aren't going to do that.
It's quite sad how Nvidia fans are all for the already fairly small PC market being segregated due to what are basically PR stunts.
Soldato
Well there is very little Physx related games about at the moment to PR with so they are using something new.
Ah yes WL WYL and Wysokie, those well known pc termonoligies?
Yes, those "well known termonoligies" are placed in the same order, just language is different. Just compare 2 screens accordingly.
Actually, most ROAM and dynamic LOD algorithms end up they costly in CPU time. If you can get hardware accelerated GPU functions to do this for you then it is easier to program and less geometry bandwidth is used sending your software calculated polygons over the bus to the GPU.
In the end it is a benchmark designed to test performances of features liek tesselation, so it makes no sense to reduce it. Do you you expect the next 3D mark/Vantage to come with rubbish graphics so older slower cards can keep up?
Maybe tessellation in this one test could be reduced without much visible impact (can we some some screen shots showing this to be true??) However, the underlying problem with ATI cards will still remain, just hidden and waiting to the next game with a lot of tessellation comes up.
What is needed is a AAA game title that heavily uses tessellation which gives large visible increases in quality but also makes use of a lot of other DX11 shaders etc., and then fair real/world comparisons can be made.
Soldato
Wrong as there is a difference to supporting a feature & supporting at specific level.
Next the only AA option in game will be 32X because both ATI & NV support AA but ATI does not support 32X so they will simply have to go without AA all together .
And how about you not being able to connect a blueray player with HDMI 3.1 to a TV with a 4.1 input just because they didn't want to support a lesser standard.
Soldato
Because it's very likely the final released game will have the same issues, are you honestly telling me it makes sense to release a game that runs so much worse on AMD hardware when they have something like 90% of the DX11 market share?
You purposefully missed the point didn't you? No one's talking about rubbish graphics, they're talking about inefficiency. Setting tessellation to 10 is completely pointless if 10 looks no different than 2.
Get that? That's theoretical, 10 levels of tessellation, 3 to 10 looking no different than 2. Not worse graphics, the same graphics, running as efficiently as possible.
Would you condone the usage of ridiculously high res textures that look no different in game to textures a 10th of the size, just because it's "high settings"? How about such a scenario where AMD graphics cards were perfectly fine with those ridiculous textures, but it crippled nVidia cards unnecessarily. It's just testing the cards after all, and if the nVidia cards couldn't cope, then it's their issue, right?
Can you please pass me this peer-reviewed and verified proof that there is no difference in image quality?
Why should NVidia users have to put up with reduced image quality just because ATI cards are inadequate at a feature they touted so much?
Maybe there is no image quality difference and the game should come with a reduced level of tessellation, or better still user selectable levels so nvidia owners can maximize the eye-candy. But that doesn't detract from the point that there is a still an underlying problem with the ATI tessellation hardware and if tessellation is used heavily in future titles then the ATI cards will perform worse.
It seems ATIs solution to the problem was a software solution based on reducing the tessellation when it has less visual impact rather than making the hardware itself faster.
If the benchmark is representative of the game then clearly Ubisoft have dropped the ball, given the volume of ATI DX11 cards out there.
I'm not biased because I own a 5870, far from it, just from a business POV I would have thought Ubisoft would want to sell as many copies of Hawx2 as possible and clearly this isnt the way to do it. Maybe the kick back from Nvidia makes it financially viable to follow this tack but I doubt it.
My gut feeling says it would be very hard to distinguish between a tesselation level that runs fine on ATI hardware and one that doesnt, happy to be proved wrong though.
I'm not biased because I own a 5870, far from it, just from a business POV I would have thought Ubisoft would want to sell as many copies of Hawx2 as possible and clearly this isnt the way to do it. Maybe the kick back from Nvidia makes it financially viable to follow this tack but I doubt it.
My gut feeling says it would be very hard to distinguish between a tesselation level that runs fine on ATI hardware and one that doesnt, happy to be proved wrong though.
Soldato
Again you miss the point in its entirety as the point is about efficiency because you will not notice the difference with the other method & everybody wins.
Maybe think also think culling is a bad idea as well & hidden & behind the field of view objects should be fully render computed just because one brand had the power to do so even though you would never notice the visual difference because its not visible in the first place.
No, I clearly understand that if there is no visible difference then it should be done that way.
But there is no evidence that a reduced tessellation level will give pixel-pixel identical results.
And you are also missing the point. If ATI's hardware can't handle high levels of tessellation then making one game that may or may not benefit from high levels of tessellation have reduced levels doesn't solve ATI's long term issues.
But there is no evidence that a reduced tessellation level will give pixel-pixel identical results.
And you are also missing the point. If ATI's hardware can't handle high levels of tessellation then making one game that may or may not benefit from high levels of tessellation have reduced levels doesn't solve ATI's long term issues.
Soldato
1) But your basing your whole argument as if there is a big enough visual difference when none have claimed it to be so.
2) its you who are missing the point that level of tessellation is not needed now or in the near future in games that will make a noticeable worth while effect & has been proven may times already in titles already.
Tessellation has to be thought out & part of the initial deign state to be effective & not tacked on after as there will be little visual difference otherwise like as been shown many times already, so they push the tessellation compute way beyond what is necessary just to make it noticeable because it has been implemented in an ineffective way.
There is a big difference in doing what can be done than doing what should be done.
And what's this about ATi long term tessellation issues as if you know more than the rest of us about what's on the map.
Last edited:
I don't think it needs to be pixel identical, it needs to have no material difference.
If you need to wheel out the screen shots, zoom in by 10x and highlight the different pixels to see it then personally I'm not bothered.
Associate
I have just run the benchmark. This might seem a silly question but how do i save the results to be posted on here for you all to see.
Do i hit the PrtScn button?? if this is how it's done nothing happens when i hit it.
Thanks.
Do i hit the PrtScn button?? if this is how it's done nothing happens when i hit it.
Thanks.
Soldato
Personally I use the built in screen grab facility in evga precision - print screen doesn't work with this bench
Using precision or similar usually saves the screen grab to a pre determined location once the allocated key is pressed
Using precision or similar usually saves the screen grab to a pre determined location once the allocated key is pressed
Soldato
I used fraps to capture screens.
Soldato
Run Fraps and press F10 when the results come up it then saves the pic in the fraps directory
Problem is... game developers need a reliable output from the source input otherwise things can break i.e. being able to see things you shouldn't because the terrain profile isn't how its intended.
Also sometimes you can't help over doing tessellation in some areas in order to make sure you get effective coverage in all areas.