ATi have just been caught cheating.

Page 4 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
All the fuss everyone is making would seem to be over nothing. ATI has replied to the issue.
<A HREF="http://www.theinquirer.net/?article=15971" target="_new">TheINQ</A>

<font color=blue>_______________________________</font color=blue>
Canada
Asus A7N8X-X, Athelon XP 2500+ Barton,
Samsung 1gb DDR400, MSI GeforceFX5900 XT.
Aquamark=<b>36077</b> 3DMark03=<b>5322</b>
 
Well it appears that Lars/Borsti got an official response from ATI which he shared with B3D. I'm sure you've read it before coming here, but just thought I'd post the page just so everyone knows.

<A HREF="http://www.beyond3d.com/forum/viewtopic.php?t=12486&postdays=0&postorder=asc&start=360" target="_new">http://www.beyond3d.com/forum/viewtopic.php?t=12486&postdays=0&postorder=asc&start=360</A>

So it is the implementation Wavy Dave was talking about. Not sure if I'm happy about it. Would like nother option in the slider to include, "not new 'intelligent' algorithm" so that you had more choice. Not looking like a cheat so much as an unwanted new feature, no application detection either. But at least if they know the effects perhaps reviewers can incorporate that in their reviews.

Anywhoo, still not happy, but not quite as dissapointed. Now just dissapointed in stupid handling of this whole thing. I expect alot more to come out of the next few days.

Now off to bed.


- You need a licence to buy a gun, but they'll sell anyone a stamp <i>(or internet account)</i> ! - <font color=green>RED </font color=green> <font color=red> GREEN</font color=red> GA to SK :evil:
 
You bastardo! I write a big long post, and you beat me by mere seconds! :tongue:

I gotta learn to read/type faster! :wink:

Shoulda responded right after Borsti's post instead of reading the whole thread. 😱

Anywhoo, I guess I'll read the InQ before going to bed.

Thanx!


- You need a licence to buy a gun, but they'll sell anyone a stamp <i>(or internet account)</i> ! - <font color=green>RED </font color=green> <font color=red> GREEN</font color=red> GA to SK :evil:
 
typical reaction to fanboys
remember the FX issue?
this is the same thing
the fanboy's perfect company had a huge flaw... they dont want to accept it

so its funny how most of the people here that say they're not biase are going
"ATi didn't cheat! it's ATi!"
when nVidia cheated it was like
"OMG I CAN'T BELIEVE IT!!! THEY'RE THE DEVIL!! nVIDIA SUX!!!"
Click to expand...
where has anyone done that at all? the most that has been said in this thread are things like "why is everyone freaking out, its not like nvidia doesnt do it already"

which is TRUE. personally, i think its really really really very crappy that ATI has done this becase like alot of people the main reason i supported them was because they didnt have to resort to IQ sacrifices to compete with the opposition..

and i disagree that this is the same thing. the FX was complete hype and lies. this is one filtering issue that will probably go away with the next drivers, now that it has been discovered *edit- read ATI's comments on this. if the IQ is the same, this is a good thing, not a bad thing*



on the other side of the coin, you could say this dude:

"nvidia fans have been bashed/flamed for the last 2 years for supporting nvidia's "cheats", and now that ATI has been caught ONCE, they are jumping on the oportunity to get revenge"


honestly i see alot of people doing that in the thread GW linked. obviously the 2nd guy that posted in this thread is happy about what happened as well because ATI is tarnished.. and how does this benifit you? nvidiots that finally have a chance to redeem their favorite companies name....... come on gimme a break, its a COMPANY that wants to make money off you

-------
<A HREF="http://www.albinoblacksheep.com/flash/you.html" target="_new">please dont click here! </A>
<A HREF="http://www.subhi.com/keyboard.jpg" target="_new">This is you, interweb junky</A>
 
Well heres my opinion. When my new x800pro arrives this week - it will rip 9800pro's and 5950's a new one - and then some.

=)

<A HREF="http://service.futuremark.com/compare?2k1=7737657" target="_new"> My rig </A>
 
This ATI issue is not that bad, if there is no application detection and there is no (or near-zero) difference in IQ why would we complain much.

When nVidia was in trouble and caught cheating, it's was obvious and the IQ was really affected, we could really tell the difference in scrrenshots.

And when this will stop?
Should we consider WHQL certication enough? Probably not.
Should Microsoft change their WHQL policy? I doubt it.
Should we blame ATI or nVidia for trying increasing FPS? Not necessarly.
Should we let ATI/nVidia increase FPS at all cost? NO.

I think there is a lot of questions and no one can answer really them to the satisfaction of everyone. Let ATI/nVidia use "innovative" technique to increase performance is good, as long as IQ on both products are comparable and not affected. Why would I complain if I can't see the difference?

I really think that "adaptive drivers" will become more and more popular in the future. This is a very good way to ensure the highest-possible performance in various situation. I really think that nVidia will benifit from this ATI leak of information. They might implement this kind of technique too.

We can compare this to F1, there is tons rules to respect. But at the end, the winner is the one that have the better driver in the seat and the better ingenieers that tweak a chassis and a 3.0Liter engine. No matter what they do, they can't have a 4.0Liter engine, larger wheels, etc...

This race to HIGH scores is good for us at the end. Both nVidia/ATI learn from each other errors and good move. At the end, the products they get to the market are very good and overall very impressive.

--
Lookin' to fill that <font color=blue>GOD</font color=blue> shape hole!
 
All the fuss everyone is making would seem to be over nothing. ATI has replied to the issue.
Click to expand...
I don't agree with you there, Vimp. Is it OK to shoot people if you say, "Yeah, I shot 'em. But it's OK because I meant to do it"?

If this method doesn't affect image quality, why was it implemented in the 9600, but it wasn't good enough for the 9800XT?

It was obviously implemented in the X800 to improve benchmark scoring, nothing more. Once again, I don't care if you have a 9600 or X800, you better be able to see identical image quality compared to a 9800 PRO with the filtering quality set to "Quality". End of story.

________________
<b>Radeon <font color=red>9500 PRO</b></font color=red> <i>(hardmodded 9500, o/c 340/310)</i>
<b>AthlonXP <font color=red>~2600+</b></font color=red> <i>(2400+ @ 2145 Mhz)</i>
<b>3dMark03: <font color=red>4,055</b>
 
Now lets add some more insanity to the mix:


By Lars:

Seems the same happens to NV40 (TrilOpt Off) and 9800XT.

<A HREF="http://www.rivastation.com/temp/atifilter2.png" target="_new">http://www.rivastation.com/temp/atifilter2.png</A>

Any ideas why this happens? Why do the colored mipmaps enhance the filtering results so much - and on all cards?

*confused*

Lars
Click to expand...
________________
<b>Radeon <font color=red>9500 PRO</b></font color=red> <i>(hardmodded 9500, o/c 340/310)</i>
<b>AthlonXP <font color=red>~2600+</b></font color=red> <i>(2400+ @ 2145 Mhz)</i>
<b>3dMark03: <font color=red>4,055</b>
 
If this method doesn't affect image quality, why was it implemented in the 9600, but it wasn't good enough for the 9800XT?
Click to expand...
Well, it does affect image quality but only very slightly, the question is whether or not it is really a perceptible shift. Of course that doesn't make things fair for benchmarking against cards doing a near 'perfect' job compared to the reference raster image.

As for it being implemented in the R9600 series and not the R9800, it may be something in the architecture that limits/inhibits it's use in the R9800 series, they are different architectures, so there may be something to that.

Also the 'benifits' of increased speed at the slight cost to IQ may have seemed more necessary on the slower R9600 series.

I'm not making excuses here, just trying to offer possible reasons. I'm sure all will be answered within the week.

I still would prefer the option to select to deactivate this algorithm.


- You need a licence to buy a gun, but they'll sell anyone a stamp <i>(or internet account)</i> ! - <font color=green>RED </font color=green> <font color=red> GREEN</font color=red> GA to SK :evil:
 
I'm going to drop out of this thread because there are enough opinions being posted. I've read all the fact's, and I'm still crying foul.

<A HREF="http://rmitz.org/AYB3.swf" target="_new">All your base are belong to us.</A>
<A HREF="http://service.futuremark.com/compare?2k3=2216718" target="_new"><b>3DMark03</b></A>
 
i agree it definately could be a bad thing that is happening


just gotta wait until people investigate it


-------
<A HREF="http://www.albinoblacksheep.com/flash/you.html" target="_new">please dont click here! </A>
<A HREF="http://www.subhi.com/keyboard.jpg" target="_new">This is you, interweb junky</A>
 
I understand, like was said many times 2 B3D I think it's going to be a more personal call than anything that can be classified objectively.

I'm not happy about this development, but for me it's not the same as some previous 'cheats' to me it's more of the order of the shader re-ordering cheat.

We'll see though. I'm still on the fence, but not happy that they didn't mention this until they were confronted with it.


- You need a licence to buy a gun, but they'll sell anyone a stamp <i>(or internet account)</i> ! - <font color=green>RED </font color=green> <font color=red> GREEN</font color=red> GA to SK :evil:
 
I don't think it's a huge deal but I don't like when either brand tries to trick consumers. This should be an option you can turn on/off.
 
I don't why some people are getting in such a huff about optimizations; the cheat thing isn't on (i.e. lying about it/only revealing under pressure, and not being able to disable it), but I don't think there isn't anything inherently wrong with optimizations.

The most important thing is IQ, and if company X can find a way of delivering more fps without lowering IQ in a way that is easily noticeable, then good for them (that is what things like texture compression and z-buffering do). Nvidia got into trouble because they sacrificed too much IQ for speed (not that it bothered everyone); it doesn't appear as if ATI have done the same.

Of course, they should be open about it and allow people to disable any new tricks if they don't like the results, but don't have a go at them (or any other company) for adopting a psychovisual approach to IQ in the way that audio/video codec people approach audio & video quality.

My brother became so obsessed with things like speaker cables and power filters etc. that he stopped listening to music. If some of you aren't careful, you will end up looking at references images all day in Photoshop instead of actually playing games.

At the moment, realtime rendering is all about compromises; it would great if gfx cards could run Mental Ray or Brazil in realtime with min 4x/max 256x sampling per pixel, but they can't. IF you need a magnifying glass to see the effect on an optimization in a screenshot, is it really such a big deal?



Axis of Stupid = coop, Kanavit, FUGGER, SoDNighthawk, and ninkey.
 
Nvidia owners remind me of the people who bought Beta video recorders so long ago, wandering aimlessly through the video store aisle of life looking for that one movie that wasn't on VHS so that they could hold it aloft to prove beta was the best, so so sad...my heart goes out to them
Click to expand...
LMAO...hey...beta hi fi rawked!!! *flips satan rules sign*

----------
<b>It is always brave to say what everyone thinks. </b> <i>Georges Duhamel</i>

TKS
 
Get some crayons and construction paper. Render everything by hand.
Click to expand...
LOL...this thread rules! *grabs big ol fat crayolas and some scratch paper* Far Cry is gonna rule you guys!!! WOOO HOOO!! Trilinear anyone? I'll just draw it three times for ya! ROFL

----------
<b>It is always brave to say what everyone thinks. </b> <i>Georges Duhamel</i>

TKS
 
Funny screenshot from Ati presentation:

<A HREF="http://home.arcor.de/quasat/ATi PDF on Benchmarking Section Texturing Bottlenecks.JPG" target="_new">http://home.arcor.de/quasat/ATi PDF on Benchmarking Section Texturing Bottlenecks.JPG</A>

If this is real, Ati's left hand isn't aware of the right hand's filtering techniques...

One final point I'd like to make:

<b>If Ati's Brilinear was as good as full trilinear, they wouldn't have to switch back to full trilinear based on mipmap color, would they?</b>

Think about it.

________________
<b>Radeon <font color=red>9500 PRO</b></font color=red> <i>(hardmodded 9500, o/c 340/310)</i>
<b>AthlonXP <font color=red>~2600+</b></font color=red> <i>(2400+ @ 2145 Mhz)</i>
<b>3dMark03: <font color=red>4,055</b>
 
Who do u think you are tellin ppl to pack there S*** up and F*** out of these forums??

You own these forums??


You a Mod??




Thought not so how about shuting up and showing ppl a bit of respect, its ppl like you who give forums a bad name.
 
My brother became so obsessed with things like speaker cables and power filters etc. that he stopped listening to music. If some of you aren't careful, you will end up looking at references images all day in Photoshop instead of actually playing games.
Click to expand...
I've had the same thought...

<font color=green>"The creative powers of English morphology are pathetic compared to what we find in other languages." (Steven Pinker, The Language Instinct)</font color=green> 😎
 
I have one small question... as this entire thread has me peeved.

Isn't it <b>BI</b>linear filtering, not <b>BRI</b>linear filtering?

Small peeve, perhaps, but it's everywhere in this thread and it seems everyone is making the same mistake. Bi means two, and Tri means three... so what the hell does Bri mean?

<font color=red> If you design software that is fool-proof, only a fool will want to use it. </font color=red>
 
Splenda has got it right. Nvidia has used "Brilinear" filtering when Trilinear was set in the driver options, and now it appears Ati is doing the same thing.

________________
<b>Radeon <font color=red>9500 PRO</b></font color=red> <i>(hardmodded 9500, o/c 340/310)</i>
<b>AthlonXP <font color=red>~2600+</b></font color=red> <i>(2400+ @ 2145 Mhz)</i>
<b>3dMark03: <font color=red>4,055</b>
 
Just noticed on THG front page that ATI is holding a live chat to explain themselves at 3pm Eastern today. Check it out <A HREF="http://www.ati.com/online/chat/" target="_new">here</A>
 
I still don't know if ATI is doing the same thing. From what they say it is different, and no one has been able to come up with a complete breakdown of exactly what is happening. The R3XX and nV3X differ in AF quality even when BOTH are using trilinear, so it may not be actual brillinear, it may simply be the oddities of this new algorithm, just like the rotational method gave anomalies under certain angles. However if it is an adaptive technique, then it might not be the same as nV but it is similar enough to warrant being called brillinear.

<A HREF="http://www.3dcenter.de/artikel/2003/10-26_a_english.php" target="_new">Here's 3Dcenter's Brillinear description</A>, they don't like the term, but it has become the established norm.

Hopefully further investigation will reveal more.


- You need a licence to buy a gun, but they'll sell anyone a stamp <i>(or internet account)</i> ! - <font color=green>RED </font color=green> <font color=red> GREEN</font color=red> GA to SK :evil:
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom