ATI's Late Response to G70 - Radeon X1800, X1600 and X1300
by Derek Wilson on October 5, 2005 11:05 AM EST- Posted in
- GPUs
High Quality AF
One of the greatest things about the newest high end hardware from NVIDIA and ATI is that advanced filtering features can be enabled at any resolution while still maintaining playable framerates. It may take developers a little while to catch up to the capabilities of the X1800 and 7800 lines, but adding value through advanced quality features is definitely a welcome path for ATI and NVIDIA to take. For this launch, ATI has improved their AA and AF implementations. We also have two brand new features: Adaptive AA and Area Anisotropic filtering.
Starting with Area Anisotropic (or high quality AF as it is called in the driver), ATI has finally brought viewing angle independent anisotropic filtering to their hardware. NVIDIA introduced this feature back in the GeForce FX days, but everyone was so caught up in the FX series' abysmal performance that not many paid attention to the fact that the FX series had better quality anisotropic filtering than anything from ATI. Yes, the performance impact was larger, but NVIDIA hardware was differentiating the Euclidean distance calculation sqrt(x^2 + y^2 + z^2) in its anisotropic filtering algorithm. Current methods (NVIDIA stopped doing the quality way) simply differentiate an approximated distance in the form of (ax + by + cz). Math buffs will realize that the differential for this approximated distance simply involves constants while the partials for Euclidean distance are less trivial. Calculating a square root is a complex task, even in hardware, which explains the lower performance of the "quality AF" equation.
Angle dependant anisotropic methods produce fine results in games with flat floors and walls, as these textures are aligned on axes that are correctly filtered. Games that allow a broader freedom of motion (such as flying/space games or top down view games like the sims) don't benefit any more from anisotropic filtering than trilinear filtering. Rotating a surface with angle dependant anisotropic filtering applied can cause noticeable and distracting flicker or texture aliasing. Thus, angle independent techniques (such as ATI's area aniso) are welcome additions to the playing field. As NVIDIA previously employed a high quality anisotropic algorithm, we hope that the introduction of this anisotropic algorithm from ATI will prompt NVIDIA to include such a feature in future hardware as well.
We sat down with the D3DAFTester to show the difference between NVIDIA and ATI hardware with and without the high quality mode enabled. Here's what we ended up with:
One of the greatest things about the newest high end hardware from NVIDIA and ATI is that advanced filtering features can be enabled at any resolution while still maintaining playable framerates. It may take developers a little while to catch up to the capabilities of the X1800 and 7800 lines, but adding value through advanced quality features is definitely a welcome path for ATI and NVIDIA to take. For this launch, ATI has improved their AA and AF implementations. We also have two brand new features: Adaptive AA and Area Anisotropic filtering.
Starting with Area Anisotropic (or high quality AF as it is called in the driver), ATI has finally brought viewing angle independent anisotropic filtering to their hardware. NVIDIA introduced this feature back in the GeForce FX days, but everyone was so caught up in the FX series' abysmal performance that not many paid attention to the fact that the FX series had better quality anisotropic filtering than anything from ATI. Yes, the performance impact was larger, but NVIDIA hardware was differentiating the Euclidean distance calculation sqrt(x^2 + y^2 + z^2) in its anisotropic filtering algorithm. Current methods (NVIDIA stopped doing the quality way) simply differentiate an approximated distance in the form of (ax + by + cz). Math buffs will realize that the differential for this approximated distance simply involves constants while the partials for Euclidean distance are less trivial. Calculating a square root is a complex task, even in hardware, which explains the lower performance of the "quality AF" equation.
Angle dependant anisotropic methods produce fine results in games with flat floors and walls, as these textures are aligned on axes that are correctly filtered. Games that allow a broader freedom of motion (such as flying/space games or top down view games like the sims) don't benefit any more from anisotropic filtering than trilinear filtering. Rotating a surface with angle dependant anisotropic filtering applied can cause noticeable and distracting flicker or texture aliasing. Thus, angle independent techniques (such as ATI's area aniso) are welcome additions to the playing field. As NVIDIA previously employed a high quality anisotropic algorithm, we hope that the introduction of this anisotropic algorithm from ATI will prompt NVIDIA to include such a feature in future hardware as well.
We sat down with the D3DAFTester to show the difference between NVIDIA and ATI hardware with and without the high quality mode enabled. Here's what we ended up with:
NVIDIA 7800 GTX AF
NVIDIA 7800 GTX AF
Mouse over to cycle images
103 Comments
View All Comments
mlittl3 - Wednesday, October 5, 2005 - link
I'll tell you how it is a win. Take a 8 less pipeline architecture, put it onto a brand new 0.90nm die shrink, clock the hell out of the thing, consume just a little more power and add all the new features like sm3.0 and you equal the competition's fastest card. This is a win. So when ATI releases 1,2,3 etc. more quad pipes, they will be even faster.I don't see anything bob. Anandtech's review was a very bad one. ALL the other sites said this was is good architecture and is on par with and a little faster than nvidia. None of those conclusions can be drawn from the confusing graphs here.
Read the comments here and you will see others agree. Good job, ATI and Nvidia for bringing us competition and equal performing cards. Now bob, go to some other sites, get a good feel for which card suits your needs, and then go buy one. :)
bob661 - Wednesday, October 5, 2005 - link
I read the other sites as well as AT. Quite frankly, I trust AT before any of the other sites because their methodology and consistancy is top notch. HardOCP didn't even test a X1800XT and if I was an avid reader of their site I'd be wondering where that review was. I guess I don't see it your way because I only look for bang for the buck not which could be better if it had this or had that. BTW, I just got some free money (no, I didn't steal it!) today so I'm going to pick up a 7800GT. :)Houdani - Wednesday, October 5, 2005 - link
One of the reasons for the card selections is due to the price of the cards -- and was stated as such. Just because ATI is calling the card "low-end" doesn't mean it should be compared with other low-end cards. If ATI prices their "low-end" card in the same range as a mid-range card, then it should rightfully be compared to those other cards which are at/near the price.But your point is well taken. I'd like to see a few more cards tossed in there.
Madellga - Wednesday, October 5, 2005 - link
Derek, I don't know if you have the time for this, but a review at other website showed a huge difference in performance at the Fear Demo. Ati was in the lead with substantial advantage for the maximum framerates, but near at minimum.http://techreport.com/reviews/2005q4/radeon-x1000/...">http://techreport.com/reviews/2005q4/radeon-x1000/...
As Fear points towards the new generation of engines, it might be worth running some numbers on it.
Also useful would be to report minimum framerates at the higher resolutions, as this relates to good gameplay experience if all goodies are cranked up.
Houdani - Wednesday, October 5, 2005 - link
Well, the review does state that the FEAR Demo greatly favors ATI, but that the actual shipping game is expected to not show such bias. Derek purposefully omitted the FEAR Demo in order to use the shipping game instead.allnighter - Wednesday, October 5, 2005 - link
Is it safe to assume that you guys might not have had enough time with these cards to do your usuall in-depth review? I'm sure you'll update for us to be able to get the full picture. I also must say that I'm missing the oc part of the review. I wanted to see how true it is taht these chips can go sky hig.> Given the fact that they had 3 re-spins it may as well be true.TinyTeeth - Wednesday, October 5, 2005 - link
...an Anandtech review.But it's a bit thin, I must say. I'm still missing overclocking results and Half-Life 2 and Battlefield 2 results. How come no hardware site has tested the cards in Battlefield 2 yet?
From my point of view, Doom III, Splinter Cell, Everquest II and Far Cry are the least interesting games out there.
Overall it's a good review as you can expect from the absolutely best hardware site there is, but I hope and expect there will be another, much larger review.
Houdani - Wednesday, October 5, 2005 - link
The best reason to continue benchmarking games which have been out for a while is because those are the games which the older GPUs were previously benched. When review sites stop using the old benchmarks, they effectively lose the history for all of the older GPU's, and therefore we lose those GPUs in the comparison.Granted, the review is welcome to re-benchmark the old GPUs using the new games ... but that would be a significant undertaking and frankly I don't see many (if any) review sites doing that.
But I will throw you this bone: While I think it's quite appropriate to use benchmarks for two years (maybe even three years), it would also be a good thing to very slowly introduce new games at a pace of one per year, and likewise drop one game per year.
mongoosesRawesome - Wednesday, October 5, 2005 - link
they have to retest whenever they use a different driver/CPU/motherboard, which is quite often. I bet they have to retest every other article or so. Its a pain in the butt, but thats why we visit and don't do the tests ourselves.Madellga - Wednesday, October 5, 2005 - link
Techreport has Battlefield 2 benchmarks, as Fear, Guild Wars and others. I liked the article, recommend that you read also.