![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
As you may know, at the time of 3DMark05's release, there have been many discussions about its validity as a benchmark... The Beyond3D forum was one of the places where such discussions were held. It was an interesting place, because there are many developers and other insightful people among the participants there.
Beyond3D basically had 3 issues with 3DMark05:
1) DST shadowmaps were enabled by default on hardware that supports it. This is against Futuremark's guideline of supporting only features that were implemented by at least 2 IHVs. It also makes apples-to-apples comparisons impossible.
2) 3Dc, another vendor-specific feature was not used, and this would also improve performance and/or quality, like DST/PCF. This makes Futuremark biased.
3) Beyond3D wasn't informed by this change of guidelines, and feels ignored as a BDP partner.
I didn't agree with any of their points, and I thought I should defend 3DMark05, because their points were simply wrong, and an influential site such as Beyond3D should be careful about giving a product a bad review based on misinformation.
In short, the reaction to these three points:
1) DST is not a vendor-specific feature. Futuremark has since revealed that there is at least one other vendor who will implement DST in the near future. So it is still according to guidelines.
As for apples-to-apples, 3DMark has not been entirely apples-to-apples in the game tests for a while now. In 3DMark03 there was a ps1.4 path in GT2 and GT3, and a fallback for ps1.1, which didn't give the exact same quality and workload, so it was not apples-to-apples. Beyond3D never complained about that either, I find it odd that they choose to make a point out of it this time.
They are game-tests, and since games aren't apples-to-apples, it is more realistic to do it this way. It gives a better idea of how games will perform on the hardware. The feature-tests are for apples-to-apples comparisons though.
2) Unlike DST, 3Dc actually still is supported by one vendor only, so it would be against Futuremark guidelines to use it. On top of that, as explained in this thread: http://www.beyond3d.com/forum/viewtopic.php?t=16710, 3Dc does not particularly suit the method of rendering that 3DMark05 uses, so whether 3Dc would be at all faster, or look better, is highly debatable. DST on the other hand, clearly gives a performance advantage, because it enables double z/stencil rate on NVIDIA hardware, and from a theoretical standpoint, the filtering gives superior quality (in practice, it appears that some people prefer the shadows on ATi while others prefer the shadows on NV).
3) It is simply Beyond3D's interpretation that Futuremark changed its guidelines. I found it rather odd that Beyond3D is the only BDP partner that complained about this. I would at least expect ATi to complain, if Beyond3D was right, since it affects ATi most. They are put at a disadvantage with NVIDIA.
As you can see here: http://www.beyond3d.com/forum/viewtopic.php?t=16643&postdays=0&postorder=asc&start=40, their interpetation of some of Futuremark's docs are up for speculation. I gave an alternative interpretation, which I believe to make at least as much sense as their interpretation, but would not lead to the conclusion that Futuremark broke their guidelines.
Anyway, why am I posting this here? Well, last week I suddenly got banned from Beyond3D, for no apparent reason. The above discussions may have been rather harsh at times, but they were almost a month ago, and I wasn't banned then.
I stopped discussing for a number of reasons. Everything had been said already, and people should just read the discussions and make up their own minds. Also, some of my posts got deleted (as you can see here: http://www.beyond3d.com/forum/viewtopic.php?t=16836, sadly you can no longer judge whether or not the post actually was disrespectful, but I suppose that was the idea), and threads got closed before people could respond to their statements.
I have posted very little since, and haven't been in harsh discussions with the Beyond3D people anymore. So why I am suddenly banned is a mystery to me.
What's more, they made a big deal out of the fact that they banned me, by creating a special thread to announce banned people (but not giving a reason): http://www.beyond3d.com/forum/viewtopic.php?t=17480
Some forum members started a discussion on why I was banned, since they seem to be as surprised as I am (http://www.beyond3d.com/forum/viewtopic.php?t=15786&start=40), but no actual reason is given, they just remain vague with things like "I've wanted to ban Scali for a while now", and "We trust you trust us in our choosing why we ban such individuals. And why posting their names may be a good idea."
My guess is that they are preparing to release their review of 3DMark05 now, and they are determined to get their points across.
In which case I would like to warn you. Beyond3D may look like a respectable 3D website, but their 3DMark05-statements are largely based on misinformation, and their motives seem questionable.
Beyond3D basically had 3 issues with 3DMark05:
1) DST shadowmaps were enabled by default on hardware that supports it. This is against Futuremark's guideline of supporting only features that were implemented by at least 2 IHVs. It also makes apples-to-apples comparisons impossible.
2) 3Dc, another vendor-specific feature was not used, and this would also improve performance and/or quality, like DST/PCF. This makes Futuremark biased.
3) Beyond3D wasn't informed by this change of guidelines, and feels ignored as a BDP partner.
I didn't agree with any of their points, and I thought I should defend 3DMark05, because their points were simply wrong, and an influential site such as Beyond3D should be careful about giving a product a bad review based on misinformation.
In short, the reaction to these three points:
1) DST is not a vendor-specific feature. Futuremark has since revealed that there is at least one other vendor who will implement DST in the near future. So it is still according to guidelines.
As for apples-to-apples, 3DMark has not been entirely apples-to-apples in the game tests for a while now. In 3DMark03 there was a ps1.4 path in GT2 and GT3, and a fallback for ps1.1, which didn't give the exact same quality and workload, so it was not apples-to-apples. Beyond3D never complained about that either, I find it odd that they choose to make a point out of it this time.
They are game-tests, and since games aren't apples-to-apples, it is more realistic to do it this way. It gives a better idea of how games will perform on the hardware. The feature-tests are for apples-to-apples comparisons though.
2) Unlike DST, 3Dc actually still is supported by one vendor only, so it would be against Futuremark guidelines to use it. On top of that, as explained in this thread: http://www.beyond3d.com/forum/viewtopic.php?t=16710, 3Dc does not particularly suit the method of rendering that 3DMark05 uses, so whether 3Dc would be at all faster, or look better, is highly debatable. DST on the other hand, clearly gives a performance advantage, because it enables double z/stencil rate on NVIDIA hardware, and from a theoretical standpoint, the filtering gives superior quality (in practice, it appears that some people prefer the shadows on ATi while others prefer the shadows on NV).
3) It is simply Beyond3D's interpretation that Futuremark changed its guidelines. I found it rather odd that Beyond3D is the only BDP partner that complained about this. I would at least expect ATi to complain, if Beyond3D was right, since it affects ATi most. They are put at a disadvantage with NVIDIA.
As you can see here: http://www.beyond3d.com/forum/viewtopic.php?t=16643&postdays=0&postorder=asc&start=40, their interpetation of some of Futuremark's docs are up for speculation. I gave an alternative interpretation, which I believe to make at least as much sense as their interpretation, but would not lead to the conclusion that Futuremark broke their guidelines.
Anyway, why am I posting this here? Well, last week I suddenly got banned from Beyond3D, for no apparent reason. The above discussions may have been rather harsh at times, but they were almost a month ago, and I wasn't banned then.
I stopped discussing for a number of reasons. Everything had been said already, and people should just read the discussions and make up their own minds. Also, some of my posts got deleted (as you can see here: http://www.beyond3d.com/forum/viewtopic.php?t=16836, sadly you can no longer judge whether or not the post actually was disrespectful, but I suppose that was the idea), and threads got closed before people could respond to their statements.
I have posted very little since, and haven't been in harsh discussions with the Beyond3D people anymore. So why I am suddenly banned is a mystery to me.
What's more, they made a big deal out of the fact that they banned me, by creating a special thread to announce banned people (but not giving a reason): http://www.beyond3d.com/forum/viewtopic.php?t=17480
Some forum members started a discussion on why I was banned, since they seem to be as surprised as I am (http://www.beyond3d.com/forum/viewtopic.php?t=15786&start=40), but no actual reason is given, they just remain vague with things like "I've wanted to ban Scali for a while now", and "We trust you trust us in our choosing why we ban such individuals. And why posting their names may be a good idea."
My guess is that they are preparing to release their review of 3DMark05 now, and they are determined to get their points across.
In which case I would like to warn you. Beyond3D may look like a respectable 3D website, but their 3DMark05-statements are largely based on misinformation, and their motives seem questionable.
welcome to real world Neo