OMG HARDOCP REVIEW OF UT2003 AND FILTERING

[whql]

Hmm. Thought provoking and plausible. I do know that Kyle had been speaking against the use of 3Dmark months before it was found out they were apparently cheating at it. I believe this point is missed throughout discussions regarding Kyle, Nvidia and 3DMark.

Given NVIDIA's responce to 3DMark03 when it was released, you don't think NVIDIA had been telling their 'Friends' what a horrific benchmark it was even before then?

 

[Blackwind]

Hmm. Thought provoking and plausible. I do know that Kyle had been speaking against the use of 3Dmark months before it was found out they were apparently cheating at it. I believe this point is missed throughout discussions regarding Kyle, Nvidia and 3DMark.

Given NVIDIA's responce to 3DMark03 when it was released, you don't think NVIDIA had been telling their 'Friends' what a horrific benchmark it was even before then?

Again, plausible. On the other hand this doesn't make it a fact.

So the UT 2003 AF issue that full trilinear can only be force on the R300 but not the NV3x isn't a cheat?

I think that would be best answered by Epic. Apparently they do not feel it is a cheat.

 

[Blackwind]

Even if you want to use that weak argument, UT2003 is a game. So there goes your argument. Even in that case he knew for over 2 months (according to him) and didn't say anything. If you felt what Kyle did is was morally correct, why did you embellish how he approached ATI? At a certain point, when the evidence is so overwhelming, reasonable people have to be willing to admit they made a mistake. Can you honestly say you are being reasonable?

Edit

And by the way he starts off by saying "We all know that Quake3 is one of the most utilized 3D benchmarks". So clearly has a problem with misleading benchmarking. Both 3DMark03 and UT2003 are very popular benchmarks.

There is nothing weak about the simple facts of the question posed. Quake is a game, 3DMark is not. It is my opinion there is a differance between the two. There isn't evidence you or anyone else can present that will change that fact. Sorry. Regarding Kyle, if you'd care to ride your Moral Pony, by all means, entertain yourself. I have better things to spend my time discussing and learning about. Rational thought is the core of one being reasonable. There is nothing unrational about my thinking.

 

[Heathen]

I do know that Kyle had been speaking against the use of 3Dmark months before it was found out they were apparently cheating at it. I believe this point is missed throughout discussions regarding Kyle, Nvidia and 3DMark.

So? Everybody's aware that OEMs choose which cards to ship in their machines based on tools such as 3dMark. The problem isn't Kyle's like or dislike of the benchmark but that Nvidia was committing fraud. Nvidia's performance in 3DMark, or practically any other benchmark or timedemo in existence, DOES NOT represent real world performance. Instead of bashing Nvidia for dishonest activities toward the consumer he just launches a personal attack against the website which raised the issue and bashed the actual benchmark. Put it this way would you ban the use of a stadium if a winning athelete was found abusing steroids during a race held there? Of course not. You'd ban the athelete.

As for UT2K3 when I ask for full-Trilinear I expect full Tri-Linear. I don't expect the driver to give me some half baked variant, regardless of the reasons why. Now if Nvidia had highlighted this partial Tri-Linear as a proper documented setting in their control panel I'd be happy as I think their hack is actually quite neat.

 

[Hanners]

There is nothing weak about the simple facts of the question posed. Quake is a game, 3DMark is not. It is my opinion there is a differance between the two. There isn't evidence you or anyone else can present that will change that fact.

UT2003 is a game, and that's what we're discussing here. I believe it's also what Fred was referring to.

 

[Fred da Roza]

There is nothing weak about the simple facts of the question posed. Quake is a game, 3DMark is not. It is my opinion there is a differance between the two. There isn't evidence you or anyone else can present that will change that fact.

UT2003 is a game, and that's what we're discussing here. I believe it's also what Fred was referring to.

Yes.

I noticed Blackwind just avoided that. Kyle ignored cheats in benchmarks and games when it involved nVidia, until of course he could give them a positive spin. Which is why I called it marketing propaganda.

As far as 3DMark03 goes, the cheats did the same thing Quack did. Gave the consumer a false impression of the cards performance. Thats the whole point.

Even if Blackwind wants to stick to games he cannot deny the double standard with UT2003. And if Kyle discovered it over 2 months ago, he must of noticed the IQ differences.

 

[John Reynolds]

I think that would be best answered by Epic. Apparently they do not feel it is a cheat.

That's because Epic doesn't care about comparative benchmarking! Were the Nvidia scores posted by themselves, the entire issue would be avoided. Throw in a competitor's scores that are derived through different, unequal settings that require more computational work and then make comparisons and it's a entirely different kettle of borscht. But you apparently refuse to see that.

I never thought the phrase apples to apples was so complex, so hard to grasp.

 

[Bolloxoid]

Again, plausible. On the other hand this doesn't make it a fact.

It is a fact. Kyle himself said he was at Nvidia discussing 3dmark03 before it was released. And surprise, surprise, after the release he publishes an article echoing Nvidia's propaganda PDF even when he himself had no understanding about the specific technical aspects. In other words, he was spoonfed.

 

[Bouncing Zabaglione Bros.]

Hmm. Thought provoking and plausible. I do know that Kyle had been speaking against the use of 3Dmark months before it was found out they were apparently cheating at it. I believe this point is missed throughout discussions regarding Kyle, Nvidia and 3DMark.

Kyle started speaking out against 3DMark when Nvidia spoon-fed him their PR document against 3DMark 2003. He quoted verbatim from the Nvidia anti-3DMark document that Nvdia released as a spoiler when 3DMark2003 was released.

Months later we found out that Nvdia's agenda and abandonment of the 3DMark2003 programme was because NV30 benched particularly badly on it, and even NV35 is uncompetative without cheats.

I tend to agree with WaltC. I've said it before - one day Kyle will wake up and realise he has been used and duped by Nvidia. I doubt he is a big enough man to admit it to himself, and will keep insisting that what he has said and done has been off his own bat, rather than from the manipulation of clever Nvidia PR staff.

 

[Bouncing Zabaglione Bros.]

There is nothing weak about the simple facts of the question posed. Quake is a game, 3DMark is not. It is my opinion there is a differance between the two. There isn't evidence you or anyone else can present that will change that fact. Sorry. Regarding Kyle, if you'd care to ride your Moral Pony, by all means, entertain yourself. I have better things to spend my time discussing and learning about. Rational thought is the core of one being reasonable. There is nothing unrational about my thinking.

So going by your previous post, you think what ATI did is wrong because Quake 3 had a loss in IQ both in game and in the timedemo? In this case (leaving aside that this was a bug that was later fixed with no loss of speed and quality), the timedemo actually represented what was shown when you played the game.

In Nvidia's case, they adjust the IQ downwards in the benchmark ONLY, so that the score that you get in the benchmark is actually much higher, and not representative of the IQ/performance that you see when playing the game.

It seems that you think an unrepresentative relationship between the timedemo and the game is okay "because it's a benchmark", but that accurately representing what you get in the game when benchmarking is some kind of cheat.

 

[kemosabe]

Again, plausible. On the other hand this doesn't make it a fact.

It is a fact. Kyle himself said he was at Nvidia discussing 3dmark03 before it was released. And surprise, surprise, after the release he publishes an article echoing Nvidia's propaganda PDF even when he himself had no understanding about the specific technical aspects. In other words, he was spoonfed.

Precisely, and this ties in with everything Walt said in his last post. Of course anyone can argue this is all coincidental, but a clear pattern is forming that suggests that [H], through the preferential treatment that has apparently been bestowed upon Kyle, is no longer at arm's length when it comes to its dealings with Nvidia. Whether they are completely founded or not, concerns about objectivity and even conflict of interest are bound to arise, and thus far Kyle's attitude has done nothing to alleviate them IMO.

 

[Blackwind]

So? Everybody's aware that OEMs choose which cards to ship in their machines based on tools such as 3dMark. The problem isn't Kyle's like or dislike of the benchmark but that Nvidia was committing fraud. Nvidia's performance in 3DMark, or practically any other benchmark or timedemo in existence, DOES NOT represent real world performance. Instead of bashing Nvidia for dishonest activities toward the consumer he just launches a personal attack against the website which raised the issue and bashed the actual benchmark. Put it this way would you ban the use of a stadium if a winning athelete was found abusing steroids during a race held there? Of course not. You'd ban the athelete.

As for UT2K3 when I ask for full-Trilinear I expect full Tri-Linear. I don't expect the driver to give me some half baked variant, regardless of the reasons why. Now if Nvidia had highlighted this partial Tri-Linear as a proper documented setting in their control panel I'd be happy as I think their hack is actually quite neat.

So if the problem is in fact that Nvidia is commiting fraud how exactly did/does Kyle have control over that? He doesn't. Whether or not he chooses to "bash" Nvidia is kind of his decision don't you think? We can question. We who support his website can most definitely question him on it. And we did. You make it sounds as though he turned around and "targeted" B3D and started bashing the benchmark. This could not be farther from the truth. I have already stated, Kyle had months before this entire situtation starting venting what he found lacking in 3DMark and where they were going. As far as the topic goes I think we are all in agreement, we want full trilinear when we want it. Simple. Not all of us are necessarily up in arms over it either. If Epic can live with is, so can I. I can only make my opinion known through what means available that I'd like control over that in the next driver release. Simple.

 

[Blackwind]

So going by your previous post, you think what ATI did is wrong because Quake 3 had a loss in IQ both in game and in the timedemo? In this case (leaving aside that this was a bug that was later fixed with no loss of speed and quality), the timedemo actually represented what was shown when you played the game.

In Nvidia's case, they adjust the IQ downwards in the benchmark ONLY, so that the score that you get in the benchmark is actually much higher, and not representative of the IQ/performance that you see when playing the game.

It seems that you think an unrepresentative relationship between the timedemo and the game is okay "because it's a benchmark", but that accurately representing what you get in the game when benchmarking is some kind of cheat.

To answer you. BOTH vendors are wrong. I have never stated otherwise. What I have stated is there is a differance between Quake and 3DMark. One is a game and one is not. Simple.

 

[kemosabe]

So if the problem is in fact that Nvidia is commiting fraud how exactly did/does Kyle have control over that? He doesn't. Whether or not he chooses to "bash" Nvidia is kind of his decision don't you think?

You are skating on thinner and thinner ice. There is no requirement to "have control" over anything - it's a matter of journalistic integrity. If you are prepared to concede that Kyle could possibly choose to say nothing to condemn such fraud while believing that it was in fact being perpetrated against the consumer, then I honestly don't think he'd want you defending him on this thread any longer.

 

[digitalwanderer]

So if the problem is in fact that Nvidia is commiting fraud how exactly did/does Kyle have control over that? He doesn't. Whether or not he chooses to "bash" Nvidia is kind of his decision don't you think

Yes but he is treating nVidia's fraud differently than he treated ATi's fraud, thus the impressions of bias and double-standards that are getting so many people irked with him and his site.

He's saying and doing two entirely different things.

 

[Blackwind]

UT2003 is a game, and that's what we're discussing here. I believe it's also what Fred was referring to.

It sure is. It also has not been addressed by either vendor as being anything but an optimization. We as consumers are the ones up in arms over it. Which is perfectly reasonable.

That's because Epic doesn't care about comparative benchmarking! Were the Nvidia scores posted by themselves, the entire issue would be avoided. Throw in a competitor's scores that are derived through different, unequal settings that require more computational work and then make comparisons and it's a entirely different kettle of borscht. But you apparently refuse to see that.

I never thought the phrase apples to apples was so complex, so hard to grasp.

I would completely disagree with you. Epic most definitely cares.

 

[Forbidden Donut]

To answer you. BOTH vendors are wrong. I have never stated otherwise. What I have stated is there is a differance between Quake and 3DMark. One is a game and one is not. Simple.

I believe he was talking about quake3 vs. ut2k3, not quake vs. 3dmark...unless I read his post incorrectly.

 

[Blackwind]

You are skating on thinner and thinner ice. There is no requirement to "have control" over anything - it's a matter of journalistic integrity. If you are prepared to concede that Kyle could possibly choose to say nothing to condemn such fraud while believing that is was in fact being perpetrated against the consumer, then I honestly don't think he'd want you defending him on this thread any longer.

LOL. Ok, do people here live their life through conjecture? Try having actually read the facts. Kyle and Brent BOTH made it very clear they were not happy with the findings of Nvidia apparently cheating in 3DMark. Never. This was not a question. They made that clear through out discussions within the forums. In fact, many of the very people here who were banned can attest to that. But this apparently wasnt good enough. If they did not cry to the heavens from the front page everyday for a week. They were lacking "journalistic integrity." That is THEIR decision to make. Not yours or mine. Simple.

It is common practice for those to judge to have all the facts presented them. In this case no one here is fit to judge IF you have not in fact, read what you believe/do not believe to be the case. The irony is that was the very reason I first came here to read B3D. This board was being accused of lacking journalistic integrity and being soley a fanperson site. I came and read for myself. I did not necessarily find that to be completely true so continued to read here.

I believe he was talking about quake3 vs. ut2k3, not quake vs. 3dmark...unless I read his post incorrectly.

You're right I misread that. What occurred with ATI and Quake 3 appears to have been an error or bug. I do not see a correlation between the two how ever.

Yes but he is treating nVidia's fraud differently than he treated ATi's fraud, thus the impressions of bias and double-standards that are getting so many people irked with him and his site.

He's saying and doing two entirely different things.

Ever consider the simple fact they are being treated different because they ARE different? Impressions are not marble inscriptions or other and very much susceptible to our own bias. Being irked does not give one the license to perform acts of character assassination simply because you disagree. I do not in any way shape or form believe Kyle acted in the correct way in addressing Rev and his post. I believe he posted very much in anger and this was not good form. I would also go as far as to say I believe he addressed Dave in very much the same fashion, in poor form. This does not on the other hand make his actions completely unwarranted. If a sentence were applied the sentence is still guilty. The "punishement" on the other hand must take all facts in account.

 

[Oblivious]

That's because Epic doesn't care about comparative benchmarking! Were the Nvidia scores posted by themselves, the entire issue would be avoided. Throw in a competitor's scores that are derived through different, unequal settings that require more computational work and then make comparisons and it's a entirely different kettle of borscht. But you apparently refuse to see that.

I never thought the phrase apples to apples was so complex, so hard to grasp.

I would completely disagree with you. Epic most definitely cares.

According to Tim Sweeney, Epic doesn't care when it comes to vendor specific issues (i.e. trilinear optimizations) or benchmarking, only driver bugs. Ideally, he would like to see this stuff handled by the API so there is no room for stuff like this, but that's not possible at this time.

http://www.beyond3d.com/forum/viewtopic.php?t=7006&sid=6ffd4ff9dd7639d71d4243da9a419825
It's the highlighted text halfway through Reverend's post.

 

[Bouncing Zabaglione Bros.]

LOL. Ok, do people here live their life through conjecture? Try having actually read the facts. Kyle and Brent BOTH made it very clear they were not happy with the findings of Nvidia apparently cheating in 3DMark. Never. This was not a question. They made that clear through out discussions within the forums. In fact, many of the very people here who were banned can attest to that. But this apparently wasnt good enough. If they did not cry to the heavens from the front page everyday for a week. They were lacking "journalistic integrity." That is THEIR decision to make. Not yours or mine. Simple.
.

Hang on, you are rewriting history here. Kyle never came out and said he was unhappy with Nvidia for cheating on benchmarks. Quite the opposite, [H] took the position that because 3DMark 2003 could be cheated, that the benchmark was now unreliable.

This is despite the fact that the same reasons given for dumping 3DMark2003 applied equally to 3DMark 2001 (which both Nvidia and [H] supported avidly). Instead, Kyle championed in-game benchmarks and timedemos, even though these are also suceptable to the same kind of cheats, and (going by Kyle's logic) should also have been abandoned for being unreliable.

Quite simply Kyle's position has flipped-flopped around, with unfathomable and tortured logic to try and justify the double standards that he appears to be applying, aways to Nvidia's benefit.