More ATI Driver News from Derek Smart

My problem with the Idea that *the 9700 Drivers Suck* *or that it was not ready for release*.. etc.. Is just plain *False*

I recieved my card today, and while I did run into a coupple little issues Overall the card is performaing as perscribed with no show stoppers. Derek recieved his board and initially found no major flaws, then with a little digging *got some*.

I dont see how any of this information equals his above cleams. I am playing a myriad of games including Sports, FPS, Adventure. heck even PSX emmulators. This card is *clearly* as ready for primetime as any other new Core released.

 

Wow, this Derek Smart thread was a fantastically entertaining read.

Anyway just to throw my newbie two cents in on this ATI driver issue. The only reason why I finally bailed on my Radeon 7500 and ponied up the $200 for a NVidia Ti4400 was a driver issue.

I could never for the life of me, get this damn board to enable VSync with the vast majority of games. I tried every single registry hack known to man with zero results. My Ti4400 works flawlessly in this regard with every single one of my games. But with my 7500, games like Morrowind would produce massive amounts of tear (using the latest Catalyst drivers).

My inner conspiracy theorist believes ATI purposefully broke VSync in order to increase benchmark scores among newbies that didn't know how to manually disable VSync when running benchmarks.

It's a shame really, since the video signal from my ATI was cleaner than the one from my current NVidia board (allowing 1600x1200 @75Hz with better sharpness), but I just couldn't tolerate the tearing in games.

This is the reason why I avoided the 8500, and why I'm avoiding the 9700. I just don't trust the drivers.

I suppose I agree with Derek that while ATI produces (for the most part consistently) superior hardware, their drivers just, well, suck.

 

Interesting...there was a similar VSYNC issue in OpenGL a long time ago that just refused to go away for some time (I think it was around the time of the TNT/TNT2. I don't remember whether or not it was solved by the time I got my GeForce DDR, which was shortly after its release...).

 

AS for proof that the 9700 is *ready for primetime*...

Look at this thread. I have pics of several games. They all look *perfect* and are playing really fast with FSAA and Aniso.

http://www.rage3d.com/board/showthread.php?s=&threadid=33635479

 

i guess its a free world right? I choose to disagree. Drivers perfect? no. Can i play all the games I am interested in withoout significant issue and with the fastest, best IQ i've ever seen? *ABSOLUTLY*

The proof is in the pudding.. and i have eatten my fill.

 

You can't prove that. For one, it's a subjective thing.

Me, I don't think errors for basic rendering operations are acceptable. Derek Smart clearly pointed out a couple of those errors.

The only time such errors are acceptable are when a particular feature is new. In the 9700, this might be an error in the implementation of displacement mapping. That I would consider an acceptable error...as long as the problem was fixed in a timely manner.

But here we're seeing a specific situation wherein multitexturing utterly breaks. That I consider totally unacceptable.

And, as Derek pointed out, neither nVidia nor Matrox have had similar errors for some time.

Then there's the fact that it is simply unfeasible to prove that every video game out there is displaying no rendering glitches on the Radeon 9700. Hell, it's impossible to do that for just one game.

Regardless, as far as I'm concerened, the fact that ATI is still having problems with errors in basic rendering functions goes to show that their driver team needs work.

 

Show us a game, appart from Dereks, where this is in error...

 

When it comes to driver issues, I think people here are confusing current /old games that work flawlessly on ATI cards with games that are still in developments. Sure... all the games are rock solid on ATI 8500/9000/9700, but the real problem are the games that developers like Derek Smart are developing on those cards. The reason why those old or current games run so flawlessly is because they have been *fixed* by both developers and drivers after it was released. Quake3 for example had a lot of bugs at first, but it was eventually fixed right??

Think about it, when was the last game that work nearly flawlessly on all videocard?? That it didn't have any bugs whatsoever?? The last game I bought was NWN, it kept crashing on me, but after a driver release and 2 patches, the game stopped crashing.... and even ran faster too.

 

I might have agreed with you RaolinDarksbane if it wasn't for the fact that BC3000AD is not new... its old been released yonks ago and it still doesnt work properly. Whose fault is it now?

 

Really? Hmmm, then there is two explanation I can think of for this:

1) ATI is out to get Derek Smart
or
2) It's his fault.

I'm leaning to 2, it takes both side(developer and driver dev) to fix this....

 

[post deleted]

 

This is the most raucous, thought provoking thread that I have seen here. I say kudos to befafactors & detractors alike! 8)

 

ah!! I love the smell of fresh napal...uhm, I mean coffee, in the morning

g'morning ladies!! Just wanted to do a quick run through before I get my day going. I'm only going to respond to the choice posts, so, don't expect too much.

Soooooo, lets get this gig going..

uhm, I don't despise him. If I did, I'd say so. My thing with Joe is that he's not particularly consistent nor clear in some of his posts. At least not that I've noticed. Thats why I stopped responding to him. And my subsequent posts about him - not TO him - are just evidence of the tongue in cheek manner in which I intend to deal with him going forward. You see, contrary to popular belief, I do have the power to ignore posts. And when I do, people can form their own opinions and kill themselves all they want. It still won't make me respond. Sooner or later, they'll learn that.

Well, I speak about it all the time. You can also check some of the media interviews I've done and if there is anything specific in this regard you'd like to ask, we should start a new topic, rather than pollute this one with off-topic banter. Here's one interesting one from CGW and here is the theme from on the cover. Ring a bell?

See what I mean? *sigh*

1. There's no implied tactic to my methods. When something irks me, I just post. m'kay?

2. I didn't come here to bitch at anyone. I've told you this before, all you have to do is READ my posts and quit making such incorrect and highly inappropriate assumptions. I came here - not knowing that ATI was even here - in response to a troll who posted crap at another board that I frequent. A board that is devoid of any IHVs or the like. Its just gamers, some media etc

3. And that nonsense about respect? Its just that NONSENSE. Here's a clue ----> PRACTICE WHAT YOU PREACH

Look, I'm not out looking for respect. Usually I'm out looking for a good debate or a sound beating. Depending on the topic, how I feel and what day it is. You don't go out seeking respect. You earn it. And if you or anyone else wants to kid themselves into thinking that respect is first and foremost on my mind, you have a LOT to learn about me. A LOT.

I have a mutual and sound respect for my peers, colleagues and people who extend the same to me. When I came here, I respectfully jumped right in, posted my thoughts and a skirmish ensued. In the, not only did YOU and your silly tinkering friends start taking my posts out of context, you felt that you had something to prove by engaging me. Nothing with debatable and/or meaningful rhetoric, but with pure adulterated BULLSHIT.

You failed to make a SINGLE point or case in contrast to what I had posted. To me, it seemed as if you were just posting for the hell of it - or maybe because you perceived yourself to be the one to take on the big bad Derek Smart and gain support therein. How's the view from over there?

Every forum on the Net has a few people who have this idea that because they are regs, there is some preconcieved alliance and/or notion of respect and the need to belong. You should get out more.

There are a LOT of people with their own views and opinions. Those who have nothing meaningful or worth processing to post, just assume an alias and just post. Others, though noted (e.g. Democoder posting a now removed link to a notorious and libelous site), just seek to cause disruption and take attention away from the issues at hand because, well, they have NOTHING meaningful to add. Its like sitting in a board room and have the wanker from PR start talking about a coffee machine, in the middle of a dev meeting. You tell him to STFU and clear the room.

Look, I know some people don't like me. However, the fact remains that NO matter what you or anyone else thinks of me, I do what I want to do. Say what I want to say. And get results. I have a LOT of say in what goes on around me and that is one thing that I am proud of. Most people kid themselves into thinking that for some odd reason, there are gutless, spineless fools, wandering the Net and with no mind of their own - and as such, just buy into all the rubbish thats posted by a bunch of creeps who think they know who Derek Smart is. They don't. PERIOD. All they want is to be like me. But alas, it takes a lot more than steamy rhetoric and libelous nonsense to get where I'm at - especially given where I've come from and what I've had to endure to get here.

So, if you're going to engage me in any meaningful dialog, make sure that the rhetoric is clear, concise and intellectually stimulating. When in doubt, DON'T POST or I'll either ignore it or engage in a manner that is unacceptable to you. Eitherway, consider yourself warned.

As I've said before, that whole mine is bigger than yours, where's yours mantra is for underachievers and untalented, gutless wannabes hiding being a keyboard and an alias on the Net. If you're going to have that stance, the best anyone can do is to prove it. After all, you can't engage in a conversation about video games is you don't play and/or develop them, can you? The same holds true for any topic. If you don't have anything to add, just read - and lets those who DO have something meaningful to add, do so. You'd be surprised at how much more you can learn from doing just that. And that, my friend, is why there are lurkers on boards, but who don't post.

You bastard!! Did you have to go and spell all that out? Couldn't u just have kept it all to yourself and let them keep guessing?

What? Thats it rubbish? You're right. It is.

Thanks for the welcome. I promise to behave

I had thought of putting in a time demo function in BC games, due to the heavy amount of processing that they do, but never did get around to doing it. I'd be willing to work with you speed kiddies on that, if you send me an email proposal indicating what you need to me to do and how. It could be a nice side project.

As for W buffer on DX9, I don't think developers have fully grasped the impact of this yet. In fact, last night I sent an email to an MS contact of mine in the DX beta program and made some specific inquiries about legacy support in DX9. Right now, its all speculation - especially since DX9 is still only at Beta 2 stage.

As for precision in 24 Z, maybe OpenGL_guy can answer that because I don't want to speculate and I do not have my 9700 tech specs on this machine. However, floating point or not, the 24bit range of the Z buffer is not enough for games with large open spaces or large worlds. We got around it - even with the W buffer - and for cards that do not have a W buffer or 24 Bit Z buffer, by doing some clever partioning along Z. It does work well, be even then, you can still see some artifacts in some circumstances. The good thing is that they don't detract from the game.

If you look at the shots I uploaded yesterday, in one shot you will see a massive station in the background with a space force marine in the foreground. If you look really, really close, you might see some artifacts on the station, though the marine is rendered just fine.

My concern with the W buffer being gone is in the interest of legacy apps. Sure, in 2004, if you are playing a game developed in 1999 and released in 2001 (e.g. Battlecruiser Millennium), you would expect to see some artifacts if you are running on hardware that only has a 24 bit Z. This is normal because you buy into the thought that, well, its legacy software and isn't expected to work right. Thats OK.

I got some ideas yesterday from DaveBaumann's contact about how to emulate (thats all it is, really) a W buffer using pixel/vertex shader. I have my doubts about it working as advertized, but it all looks good on paper. If that works, its no biggie, because after all, thats some of the things we expected from a shader engine. Then you come to the problem where, what if the card doesn't have a pixel shader part? OK, so, as before, we're back writing exclusion code for who's hardware? ATI. See where this is going? The more things change, the more they stay the same.

IMO, the lack of a W buffer in DX9 (which is not yet finalized) is NOT a good reason to remove if from the hardware. In this case, the 9xxx series.
If we can deal with some artifacts in hardware, as we have in the past and all these years, why didnt' ATI leave us to deal with artifacts (as they claim) in using a W buffer on the 9xxx series? Why make the decision to remove it and, get this, NOT TELL ANYONE. I have search the entire ATI dev forum, including the secure sites that I have access to and there isn't a single word about this. At least not that I could find. I even check the DX9 Beta docs in case I missed something. The first time I found out about this, as I stated earlier, was when I noticed artifacts that I attributed to W/Z artifacts, peeked as the card's CAPS and yelped.

You know, this is NOT the kind of thing you just quietly remove and not tell anyone. OK, breaking hardware TnL or an aspect of MT, is one thing - but flat out removing a tech that quite a few games rely on, is just bad. Very bad.

LOL!! Talk about lamb to the slaughter.

No seriously, the BC suite of games are NOT for every one. If you are into deep, HUGE and highly involving games - complete with a learning
curve, you're in for a treat. If you're looking for a quick fix, look elsewhere.

Take my advice, read the reviews first, then try the Episode 2 demo (the one with the space and planetary regions and which runs on DX8/9) before you fork out your $19.99 (it has been reduced, now that its almost a year old) right here.

You should also check out the forum threads, particularly this one which deals with all the crazy and whacky things people do in my game.

AS proof that the 9700 is *not ready* for primetime.....

Look at this thread. All 18 pages of it. They all do *not* look perfect and are mostly *not* playing fast, really fast or otherwise. In fact, in most cases, they're not playing. PERIOD. Some got going after ATIs quick patch released on 9/10. The card was released more than six weeks ago.

Enjoy!

http://www.rage3d.com/board/showthread.php?s=&threadid=33633219&perpage=20&pagenumber=1

Rubbish. BCM and BCG have the SAME problems on ATI boards.

For one thing. Here is the release timeline for the BC games. And only the FIRST one was controversial, as a result of it being released in Beta by Take2

1996 - BC3K, Take 2 Interactive / Gametek (UK)

1998 - BC3K v2.0x, Interplay

2001 - BCM, 3000AD Inc

2003 - BCG, 3000ad Inc

Get your FACTS right! Is that so hard? Have you even read ANY of the reviews of BC3K v2.0x or even BCM? No? Then go get educated. You might learn something.

It was fun back then, right now it has degenerated to a mass of dangerously litigious and libelous nonsense in aid of an orchestrated character assassination. Some people must mistake me for Bill Gates.

 

Derek,

Yes the 9700 has had some issues with some games but as you can see ATI is in that thread and getting the bugs recorded and fixed, alot of the issues are cheap powersupply related...and some have not even installed the drivers properly, some Motherboard locks up with fastwrites enabled some people have lingering previous drivers.

Rage3D is a tough one to determine a cards stability, Rage3D is the premier ATI forum...there is no other with that many members while other cards issues are spread over 12-13 sites but they are there...its always been that way.
For every link that you provide a ATI problem I can show you some for other IHV's, they are just spread all over the place..in fact some of these issues my not affect you as they are hardware related etc..

For example:

Rage3d has:
22,785 members

Nvnews has:
1,509



So of course alot of posts will appear at Rage3D with issues, but for members that have a history of building Pc's like this example..no issues:

http://www.rage3d.com/board/showthread.php?s=&threadid=33634595

 

How many existing games use a W-buffer? Since Mr. Smart has implied that there are others, it would follow that the only way he knows that is if he also knows which ones.

I'm just curious as to which ones... that's all.

 

Sure.

And hopefully, you'll provide some evidence for what irks you (as you did), and if people have disagreements over whether or not that evidence is valid or otherwise justifies some of your comments (which they did), they you wouldn't go into a tirade over it (which you did), but continue to explain your views rationally.

Fine, and immediately your posts needed editing because of the over-the-top tone, etc.

I do.

As you said later on, respect is not granted or "sought" its earned. And no, being a developer does not earn you respect on a message board. The quality of your posts do.

If you're ever "looking for a sound beating" please do it elsewhere. Quality debates are a great way to educate everone. Flaming rants aren't. Whether or not you are looking for respect is irrelevant. Having respect for others is. Though you will find that if you choose quality posting instead of endless ranting, you'll find that respect will follow, whether you're looking for it or not.

Yes, I challenged your stance and your evidence on how "crappy" the ATI drivers and the dev team is, and your insistence that they are hurting consumers.

Hmmmm...that's a stark change in contrast from your "Some good points all around. " in your first response to me. Consistency in your thoughts and posts is also something that leads to respectabiliy, not that you care of course.

 

The lack of Dx 9 being here for this card, and the of the final lack AGP 3.0 specification (Intel just released the final AGP 3.0 spec yesterday) , yet SIS and VIA have AGP 8x boards out already and are some of issues that are giving 9700 owners grief...without final spec how did these two chipset companies make AGP 8X chipsets already

http://mirror.ati.com/companyinfo/press/2002/4534.html

 

I agree with you on all points. Can I go now?

 

You know, I've been thinking about that one.

HOWEVER, it is not uncommon for IHVs to have access to prelim proposals and/or near-final specs before the final announcements are made.

This would be the problem in this case. The fact that ATI has to recall their boards, due to this 8x issue, indicates that the problem was probably on their side (OK, there's a surprise) and probably down to lack of communication or *gasp* the board was clearly not ready for primetime. Maybe something in the 8x spec changed, but ATI banked on that not being the case and shipped the card anyway. This, of course, is all speculation. But think about it. Why aren't the mobo manufacturers not the ones doing the recall?

Dunno why ATI is doing the recall anyway, except maybe to satisfy the consumer advocates and not face the threat of a class action suit for false advertizing. The card's not even going to be able (the last time I checked) to do a full 8x support atm anyway. And this problem can easily be worked around - I think - by setting the BIOS of those boards to 4x

FWIW, its kudos to them for the recall, but I'm not handing out any brownie points for that. Why? Because I don't see where they have a choice to just ignore the problem.

 

Your penchant for taking posts out of context, shines through once again.

In fact, let me refer you to my FIRST post to you. Read it. Absorb it. Think about it. THEN you'll have your answer.