Rage3D Discussion Area

Rage3D Discussion Area (http://www.rage3d.com/board/index.php)
-   AMD Radeon Discussion and Support (http://www.rage3d.com/board/forumdisplay.php?f=45)
-   -   Rage3D Reviews HIS' HD 5870 iCooler V Turbo Video Card (http://www.rage3d.com/board/showthread.php?t=33963803)

Lupine May 24, 2010 01:06 PM

Rage3D Reviews HIS' HD 5870 iCooler V Turbo Video Card
 
AMD's Adaptive Anti-Aliasing feels the heat as we test the HIS ATI Radeon HD 5870 iCooler V Turbo. Check out the performance, overclocking and crossfire scaling at 1080 and Eyefinity 3 resolutions in our latest review.

HIS Radeon HD 5870 iCooler V Turbo Video Card Review

demo May 24, 2010 08:35 PM

Great review except you are using Adaptive AA in BC2, Dirt2, Stalker CoP and Heaven demo-DX11

I cannot (or anyone else) get adaptive AA to work in DX10/11, are you sure its working? and if its not dont you think those results and the first sentance of your review are a bit misleading?

Or are you running those games in DX9 mode? although your Heaven bench specificaly states that it is in DX11, what gives? are you using a newer driver which fixes AAA in DX10/11?

and this quote -

" In these tests Adaptive Anti-Aliasing was used. Adaptive MSAA is Transparency AA that effectively takes extra samples in areas that might need it, as well as improving AA on transparent textures (like the good ol' chain link fence), improving image quality. If you're buying a HD 5870, you're interested in high quality images as well as high performance - likely you'll want this mode enabled, a lot."

Well I would like to use it if it actually worked.. am I missing something?

sorry for the confusion, thanks.

noko May 24, 2010 08:49 PM

I thought adaptive AA works in DX10/11, it is SSAA that does not work in DX10/11 from the control panel. ???

demo May 24, 2010 09:01 PM

They are both the same except SSAA super samples Alpha textures, it is not real full screen SSAA as the name would seem to imply but the method of AAA being used.

And no, MSAAA and SSAA both do not work in DX10/11, load up BC2 and have a look for yourself.

Dirt2 & Heaven demo (and Stalker CoP if enabled via console) have Transparency AA built into the game and activates when you enable AA regardless of AAA settings in CCC.

noko May 24, 2010 09:15 PM

The scalling of CF-X was awesome in this review! Looks like a 6 core processor may help there or is it just better drivers now? The card is kinda expensive but then if one is going to bye MW2 at full price it would seem justified, if one already had MW2 I would pass on this card but then that is me unless one can sell the game for $35-$40 buck. Wished game bundles had several options to pick from, what if one picked up two of these cards? I guess selling one copy or both MW2 would be an option.

As for AA settings if it is not available for DX10/11 titles unless the game automatically uses it maybe ATi should prompt the user DX9 settings only or something. I really thought I did a number of bendchmarks with MSAA and Adaptive and got different frame rates with DirtII, will have to retest now. Good thing to know if true.

caveman-jim May 24, 2010 09:43 PM

Adaptive AA does less work in titles that have their own transparency AA method, but the Cat AI profile does a little extra for where the in-game specified transparency and alpha's aren't specified. It's a subtle difference that won't be noticed unless you're really hunting for it, but I thought it more accurately represented how people use CCC - set their quality settings, close it up and go until they have a problem (for example, no AA/AF in UT3 so have to force it).

demo May 24, 2010 10:51 PM

Sorry Jim I dont fully understand what you are saying, error on my part that is for sure :P

In games that have transparency aa built in, the driver does a little "extra work" for better image??

And in DX10/11 games with no built in transparency AA there is none what so ever??

And I'd still like to know about the benchmarks, BC2, Stalker CoP, etc. did you guys have working Adaptive AA in those titles like the bench description seems to imply?

caveman-jim May 25, 2010 08:48 AM

Adaptive AA does several things; as well as offering Transparency is improves MSAA with extra samples and optimizes the image/performance depending on the profile. Where in-game options exist the setting enhances those options, not overrides.

demo May 25, 2010 10:51 AM

Ok, thanks for the explanation.

Any chance they will add an override function to the drivers so we can force Transparency AA in all titles? Considering a lot of titles from now on will have DX10/11 support I feel this is sorely needed.

caveman-jim May 26, 2010 11:01 PM

Quote:

Originally Posted by demowhc (Post 1336262934)
Ok, thanks for the explanation.

Any chance they will add an override function to the drivers so we can force Transparency AA in all titles? Considering a lot of titles from now on will have DX10/11 support I feel this is sorely needed.

Not in the near future as it requires a fundamental change in the driver architecture. I have advised it might be in their best interests to do so, but the problem is very few people have enough of an issue with games image quality, that is not due to the developer deliberately taking short cuts or optimizations. Personally I would rather AMD worked with the developers to get better MSAA + Transparency implementations for all hardware in the game at release, rather than trying to kludge it with driver fixes after the fact.

Quote:

Originally Posted by osoipl (Post 1336263758)
Great review, thanks.

You're welcome, thanks for reading!

acroig May 29, 2010 11:06 AM

Jim:

The Bad Company 2 chart on page 8, I'm curious about the minimum FPS on the Sapphire card. Any explanation for the discrepancy?

caveman-jim May 29, 2010 11:20 AM

Benchmark variation, it was a single frame but I don't like to repeat benchmarks to cherry pick best results, instead use the result as an example of actual performance. I'll repeat a bench if average is way off, but as you can see the highs and average are in line with the rest of the results showing that single low didn't impact the gameplay in any way.

acroig May 29, 2010 11:27 AM

Quote:

Originally Posted by caveman-jim (Post 1336267377)
Benchmark variation, it was a single frame but I don't like to repeat benchmarks to cherry pick best results, instead use the result as an example of actual performance. I'll repeat a bench if average is way off, but as you can see the highs and average are in line with the rest of the results showing that single low didn't impact the gameplay in any way.

:up:


All times are GMT -5. The time now is 01:21 AM.

Powered by vBulletin® Version 3.6.5
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.
All trademarks used are properties of their respective owners. Copyright 1998-2011 Rage3D.com