SimHQ - Bubba does the ATI 9700 Pro.......

Status
Not open for further replies.
Huh?....did someone mention my name?.....sorry, been playing some nice games on my 9700 with the totally SUCKY drivers......at 6X FSAA & 16X anisotropic filtering......games like UT2003 Demo, NWN, Ashron's Call 2 Demo, NOLF Demo, Morrowind, Dungeon Seige, IL2 Sturmovik, Counter Strike(no fsaa), DAOC(no fsaa), Motoracer 3, Crimson Skies, CFS2 and Serious Sam 2............well, ya know, doesn't look like I'm playing anything by Mr. Smart.........Hmmmm.....everything I try works.....must be those SUCKY drivers..... BTW, how much does an african swallow weigh? :rolleyes:
 
A Z buffer "wastes" a lot of precision for close up objects, whereas a W buffer spreads it out more evenly. It's essentially the same as a Z buffer, except you're not storing Z/W , you're storing W. It allows for a more linear depth buffer. I don't know what DS is talking about. 24bit z works just fine on my 9700.
 
I know.....but I didn't want to give Mr. Smart something he could use to explain how it's a "bug" in the drivers....... :rolleyes:
 
I'm not an activist, and I'm not a cynic. I'm just a person who wants to introduce an important, but underrepresented, angle on Mr. Derek Smart, Ph.D.'s headstrong campaigns. First, the misinformation: Derek suggests that everyone with a different set of beliefs from his is going to get a one-way ticket to Hell. Where the heck did he come up with that? In other words, what happened to his common sense? Well, while you're deliberating over that, let me ask you another question: In view of his malicious nostrums, what does it make sense for us to do now? Now, not to bombard you with too many questions, but inasmuch as I disagree with his accusations and find his ad hominem attacks offensive, I am happy to meet his speech with more speech and, if necessary, continue this discussion until the truth shines.

You don't have to say anything specifically about him for him to start attacking you. All you have to do is dare to imply that I should focus on concrete facts, on hard news, on analyzing and interpreting what's happening in the world. Perhaps Derek has some sound arguments on his side, but if so, he's keeping them well hidden; all the arguments I've heard from him are entirely footling. The older he gets, the more hateful he becomes. Not that I've come to expect any better from him. Even if one is opposed to puerile fetishism (and I, not being one of the many grotty, randy spoilsports of this world, am), then surely, if I said that the best way to make a point is with foaming-at-the-mouth rhetoric and letters filled primarily with exclamation points, I'd be a liar. But I'd be being completely honest if I said that it's unctuous for Derek to make teetotalism socially acceptable. Or perhaps I should say, it's self-pitying. Derek often expresses great interest in, and approval of, violent acts reported in the press -- spousal abuse, shooting sprees, capital punishment, and so forth -- which is another way of saying that Derek would have us believe that he is a perpetual victim of injustice. That, of course, is nonsense, total nonsense. But Derek is surrounded by stinking apostates who parrot the same nonsense, which is why his grunts would sooner ally with evil than oppose it. I'll say that again, because I want it to sink in: He cannot completely conceal his true animus and inspiration. He seeks scapegoats for his own shortcomings by blaming the easiest target he can find, that is, the most silly lunkheads you'll ever see.

Derek uses the word "pharmacodynamic" without ever having taken the time to look it up in the dictionary. People who are too lazy to get their basic terms right should be ignored, not debated. I discussed this topic in a previous letter, so I will not go into great detail now, but if you're not part of the solution, then you're part of the problem. Derek can write anything he wants about how things would be different were we to give into his demands and let him guarantee the destruction of anything that looks like a vital community, but this makes me fearful that I might someday find myself in the crosshairs of his politically incorrect claims. (To be honest, though, it wouldn't be the first time.) Let me end by appealing to our collective sense of humanity: Mr. Derek Smart, Ph.D.'s goal is not to oppose nerdy blackguardism but to reinvigorate it with a self-aggrandizing new purpose.

;)
 
Maybe I will show some ignorance here, well yep that is why I am asking. Anyways a depth buffer (Z) at 24 bits:

2^24 = 16,777,216 levels or depths of precision right? For each pixel on the screen? But is non-linear allocating more bits for objects near then far, right?

This is what Nvidia says about the W-Buffer
W-Buffering in Direct3D
Doug Rogers
NVIDIA Corporation
drogers@nvidia.com
"W-buffering is a depth-buffering alternative to z-buffering, and should be used in cases where
z-buffering produces artifacts. W-buffering does a much better job of quantizing the depth
buffer." [D3DIM.DOC]
W-buffering provides a linear representation of distance in the depth buffer. Z-buffering is nonlinear
and allocates more bits for surface that are close to the eyepoint and less bits farther away. . .

http://nvidia.com/docs/IO/1339/ATT/W_buffering2.pdf

So why would Microsoft drop W-buffer support in DX9? Seems like a better way to ensure objects are correctly rendered.
 
Thing is, you DO want more precision close up. Z fighting is far more noticable close up than far away. I want more precision at 1 meter distance than I do at 1000 meters. Also with 24/32 bit z buffer, the z fighting issues close to the far plane are less pronounced. Using a W buffer also requires a special projection matrix. This matrix doesn't work the same way as a matrix for projected textures would work (for example), which means that you'll generate wrong/different depth values for your projected texture pass. Things like depth textures would be screwed too, I think. Someone will correct me if I'm wrong I hope :) It's been a while since I used a w buffer *cough*Voodoo*cough*.
 
Well if I was Derek I would be mad as hell as well, why?

Well being a business man and having a product that I can sell will keep me in business, my employees happy not to mention all our wives. Plus the oppuntunity to move on to bigger and better and more fun projects which hopefully makes it all worth it. Now if I have to keep going back to my previous projects and messing with them to keep them working will cost me money, time and make my life probably alot less enjoyable.

So what did ATI do:

  • Dropped the W-buffer support in hardware without an easy workaround available.
    Dropped down the Z-buffer to only 24bits precision
    Did not warn the developers or inform them of the changes made nor the reasons why until after hardware was launched

Well Derek games require hardware to work upon.
The games where developed around known hardware standards and API standards which for the most part are suppose to be standardized and not break latter on when the API is updated or hardware is updated (at least not in the forseeable future).

I see ATI not being too developer friendly in this matter, Derek is a developer which was affected which will cost him money. So he isn't being anti ATI but opposes what ATI did or how they did it without helping him out (well at least that how I see it). Well I don't speak for Derek but I am glad he isn't afraid to speak up.

Now I really am more interested in if the W-Buffer is going to be replaced by something else in DX9 and what it is?

Thanks Fresh for helping me out in understanding W-Buffers and explaining it so that I can understand :).
 
Sure you can work around it : Use a Z buffer like everyone else. And FYI : NVidia doesn't support 32 bit z either. It's 24 z /8 stencil. There's a flag in D3D which tells you whether or not the card supports W buffer. It's the game's responsibility to query for this, and if it's not supported to have an alternative. Just like you'd expect a game to downsample their 2048x2048 textures if your hardware only supports 1024x1024.
 
Derek's business problems are not related to the W-buffer. Writing a game that only appeals to a very small niche market and releasing it with exponentially more bugs than ATI's drivers I'd say was a big factor.
 
Hmm, sorry if I go off topic a bit from the usual tirade of "ATI drivers suck arse" and 'No they dont, they rule' for a moment..

Derek-
Well, I'm a simmer - and he's playing the same sim that I am. There IS a speed degradation in dense scenes, banking etc. Send him email and ask him why he didnt' bother to note that. I'll tell you his answer if you'd like : the loss is neglible.

Something that's been bothering me about the 'speed degradation' that you reference- and it seems pretty specific to the 9700 Pro, and also led to my upgrading several machines to XP Pro + SP1...

In games where performance suffers when panning view into dense scenes, by simply setting the game to "Compatibility mode" (change the properties of the shortcut) and selecting "Windows 98/ME" the problem is entirely removed. Now, I can't speak for the sim referenced, but would be interested in your findings.

I did try a parallel test with one of the most problemed titles I have that illustrates this (Everquest in the Bazaar zone with 330+ player models) and the framerate on both WinXP and 98se was around 7-13fps w/ fraps panning view towards dense cluster of models.

Enabling "Compatibility Mode" on XP Pro for "Windows 98/ME" and suddenly jumped the framerate to 23-33 fps on the same exact pan. Obviously this is a worse-case example, but it's an improvement well over the same hardware with *native* Win98se/ME (which still encounters the speed degradation).

I'm unsure what this compatibility mode actually does, but would be interested if any others are able to experience noticeable improvement. It might go a long way towards cracking the oyster (so to speak) of what is the source of the degradation is, and has become a favorite "trick" of mine to remove a heavy feel to modern/post modern DX titles.
 
Well Fresh from what I understand (not much) Derek seemed to have used the right method for his game with the W-Buffer, where he has a rather large 3d space with objects throughout the space vice a bunch of close up objects. So the accuracy of far objects and not so much accuracy of a few near objects seems to make sense due to the nature of his game.
I didn't know that the W-Buffer wasn't supported in DX6 capable hardware and above. Meaning I thought that was a hardware standard that was pretty much used by all recent graphic chips.

As for the bugs in his games I wouldn't know since I can't seem to find any of his games at the store here.
 
noko said:
Well Fresh from what I understand (not much) Derek seemed to have used the right method for his game with the W-Buffer, where he has a rather large 3d space with objects throughout the space vice a bunch of close up objects. So the accuracy of far objects and not so much accuracy of a few near objects seems to make sense due to the nature of his game.
I didn't know that the W-Buffer wasn't supported in DX6 capable hardware and above. Meaning I thought that was a hardware standard that was pretty much used by all recent graphic chips.

I don't know what his code looks like. He's PROBABLY messing with the projection matrix quite a bit in order to be able to model his universe properly. You'll always run into precision problems when trying to do what DS is doing. I'd like to point out that with a z buffer the far plane doesn't even matter much at all. In fact I put my far plane to infinity. It's all in the near plane, and DS mentioned he "seeds" it to 0.001 which is ridiculously close. In most of our games we use 1.0 (feet). If I were to model a universe I'd render the scene in multiple passes (for close and far objects) using several z buffer clears combined with appropriate projection matrices.


As for the bugs in his games I wouldn't know since I can't seem to find any of his games at the store here.

Now there's a surprise :rolleyes:
 
Derek Smart [3000AD said:
]
Whats your point? His review indicated that he was using the 775 drivers which ship with the card...
The intended point was to see how people felt according to which drivers should be used in a review. The ones with the card? latest official? or latest beta?

Just a question, I wasn't prodding. :)

Derek said:
While I too like the review, I have no clue what exactly it puts into a different perspective...
Reviewing with bias on the genre of sims is what I found to be different, that's all.
 
Let's try this quoting out of context business...

Derek Smart [3000AD said:
]So, sue me for voicing MY OPINIONS.

Are they really YOUR opinions?

DaveBaumann said:
Who’s spinning Derek? Where is the spin?

I predict we'll see more of this until a certain IHV release... I wonder if a #1 dev line was enough carrot...;)
 
[quote="Derek Smart [3000AD


The 9xxx series of drivers - IMO - are just as buggy as the 8xxx series of drivers. There is no inconsistency in this regard. None whatsoever.

While others - including myself - are touting ATI's improved driver support, I still maintain that they have a bloody long way to go.
[/quote]

This is the only thing that Derek has said that I can't parse. On their face those two statements, which appeared consecutively just as I've shown them, seem to contradict each other. Possibly I misunderstood the context.

Y'know, as an 8500 owner (and a RadLE and Rage Fury owner before that), I really don't mind at all having a Derek Smart pounding on ATI for better drivers. I *would* mind if the *only* thing he was doing was pissing and moaning, but it isn't. There have been quite a few examples in his various posts where he has quoted his various communications to ATI and it is clear (to me at least) that those communications are helpful, polite, and even good humored. Helpful in not just pointing out a problem, but in also (at least the ones I saw) suggesting where the problem may be.

Derek must be an honorary Israeli tho --he seems to believe in massive retaliation. The threads I've seen go to hell start with Derek making a strong, but not unreasoned statement, then somebody makes it personal, and then there are smoking craters all over the place. . .
 
Status
Not open for further replies.
Back
Top