Catalyst 2.5 impressions

[Chalnoth]

Went ahead and installed these drivers. Overall, the result was underwhelming. I went through and took a look at the problems I was having previously:

Stuttering in NWN and Tenebrae is gone now, though the performance still isn't as high as it was on my GeForce4 with "adjusted" settings (That is, these games take more of a performance hit with FSAA enabled than does a lower level of FSAA on my GeForce4). Doesn't mean it's lower performance than the GeForce4, just not nearly as much higher as other games.

The "shiny water" doesn't draw at all in Neverwinter Nights (They just enabled Radeon support for the shiny water in the latest patch...doesn't work for me yet, though...don't know if it works on older drivers). I also noticed some glaring texture aliasing with anisotropic enabled on a particular texture in the game.

Tenebrae had some rather significant graphical glitches (looked like texture corruption, though always in the exact same place in the menu/loading screens). HUD turns grey.

Unreal Tournament: S3TC no longer works with Vogel's OpenGLdrv.dll (I'm pretty sure this worked fine in previous drivers), can't force-on FSAA in D3D mode.

Pretty much everything else is the same.

[-=DVS=-]

Just from curiosity on what are you running these drivers LoL :D any system spec :confused: do you have a computer at all :p Just kidding :p

[Fotis]

Yes shiny water is broken in 2.5 but Bioware will release a patch in a few days that fixes it.I think the problem with ATI is that they don't pay attention to game devs.They should work with them before a game is released and not after.

[AGP64]

Originally posted by Fotis
Yes shiny water is broken in 2.5 but Bioware will release a patch in a few days that fixes it.I think the problem with ATI is that they don't pay attention to game devs.They should work with them before a game is released and not after.

This seems more a PR/marketing flaw of Ati than their drivers :p
Most games were developed on Nvidia hardware until Ati took over the performance crown.

Now that Ati is becoming more and more industry leader and have introduced the Mojo day to convince developers of the benefits of using Ati hardware for games development I have good hopes that issue will resolve itself in the (near) future :)

[Chalnoth]

Originally posted by AGP64
This seems more a PR/marketing flaw of Ati than their drivers :p

Not if it works on a Radeon 8500. Anybody know if the shiny water currently works on an 8500?

[AGP64]

Originally posted by Chalnoth
Not if it works on a Radeon 8500. Anybody know if the shiny water currently works on an 8500?

post of Hellbinder on rage3d:

You know...
The new drivers Work great for me... as have the last 3 driver releases. My only issue today is shiny water in NWN, and that I have been told by An ATi employee is a bug in game.. not the driver. And will be fixed in the NWN 1.28 patch.

See following thread: http://www.rage3d.com/board/showthread.php?s=&threadid=33652490

I hope this answers your question
:D

[Chalnoth]

Well, it just seems strange to me that they (Bioware) would release something that definitely didn't work. It had to work on some machine somewhere, I should think. While it could still be sloppy coding on Bioware's part, it just seems obvious that poor drivers are the main reason why it doesn't work now (Ex. overly-lenient drivers allowed it to work in some other driver set, but not now...). Unless, of course, the person who coded it at Bioware didn't actually test the code...but that would seem like a dumb thing to do....

[Joe DeFuria]

Well, Chalnoth,

One would think a discussion on Catalyst drivers and your issues might be better served on Rage3D, don't you think?

In any case, concerning NWN, Sirec said (paraphrasing) the following:

The current shiny water patch by bioware should not work because the code is bad. However, previous ATI drivers were forgiving of the code and hence, it worked.

The new drivers are not forgiving (and they shouldn't be). So bioware is fixing their code, and will hopefully release it with the next patch.

Now, consider that this happens between driver revisions on one card. Think of what must happen between two different platforms like GeForce and Radeon. GeForce can have whatever "forgiving" code in its drivers, that doesn't get noticed because devs have primarily used nVidia hardware the past couple years. Then if you run on hardware that is not forgiving of bad code....it breaks.

Same thing happened in the 3dfx to nVidia "transition".

There's no easy way around this problem. Ideally, whoever is making "forgiving" drivers would make them more rigid to promote proper and more compatible code. But then that would probably break a lot of existing code, so of course IHVs are reluctant to do that.

In any case, I'm glad that ATI is willing to "break" compatibility for the sake of proper operation. And with ATI now having the most advanced hardware on the market, devs will be using more than one platform for coding...

[Chalnoth]

Originally posted by Joe DeFuria
The current shiny water patch by bioware should not work because the code is bad. However, previous ATI drivers were forgiving of the code and hence, it worked.

Which is what I elluded to in my previous post. It's still bad driver coding on ATI's part, though it is good that they fixed it.

And if you're going to want to try to accuse nVidia of having poor drivers that cause ATI's problems, you're going to need to show very specific examples. All of the developers I've seen give specific examples have said that nVidia's drivers are better.

By contrast, I've seen some developers state that 3dfx's drivers were very poor, and Matrox's drivers are the best of any company.

[Bigus Dickus]

Originally posted by Chalnoth
Well, it just seems strange to me that they (Bioware) would release something that definitely didn't work. It had to work on some machine somewhere, I should think. While it could still be sloppy coding on Bioware's part, it just seems obvious that poor drivers are the main reason why it doesn't work now.

You're racking up on naive statements for the day. Strange that any company would ever release code that didn't work, huh? But here you are arguing that while it's "OBVIOUS" that one company released bad code, it seems strange to you that another would.

Jesus. The things some people will say to defend their opinions.

[nin_fragile14]

The new drivers work great for me. The last two sets haven't worked too well, so I'm glad these are fine. The V-Sync options are a nice feature that has finally beed added.

[jbirney]

Originally posted by Chalnoth
Well, it just seems strange to me that they (Bioware) would release something that definitely didn't work.

But they did. When NWN shipped there was no shinny water on any cards but Gf3/4. why? Sloppy drivers? No. Its because the folks at Bioware chose to use custom NV openGL calls. That alone should raise a flag.

[AGP64]

Another nice example that Ati drivers are becoming more mature with every release

Having always used Nvidia until 2 weeks ago when I got my Sapphire 9700 Pro I can say I am extremly pleased so far. These drivers on the surface do not seem any better or worse than the last to me. For what its worth 3dMark is up from 13,521 to 13,741. I did a fresh install with this new card & removed old drivers by this guide http://www.driverheaven.net/drivers...?page=uninstall
& have had zero problems & I have a Via chipset.

To have input from a ATI Engiineer is pretty cool , thanks!


from the following thread on rage3d http://www.rage3d.com/board/showthread.php?s=&postid=1331590967#post1331590967

[Chalnoth]

Originally posted by Bigus Dickus
You're racking up on naive statements for the day. Strange that any company would ever release code that didn't work, huh? But here you are arguing that while it's "OBVIOUS" that one company released bad code, it seems strange to you that another would.

Jesus. The things some people will say to defend their opinions.

You don't seem to understand. I meant to say that there is an ATI card out there somewhere, with some driver version, that should most certainly work with that code. In other words, it's only not working now because ATI's drivers in the past (Or with another card) were too lenient, apparently (coupled with sloppy coding, but if it works, why bother to troubleshoot it? But...apparently somebody didn't do enough testing...).

[ATI LoVeR 9700]

HOLY ****!

THIS IS A ******* nVIDIA OPTIMIZED GAME! GRRR!!! YOU EXPECT IT TO WORK PERFECT ON ATI HARDWARE?!

WHY SHOULD ATI HAVE TO BEND BACKWARD TO GET IT TO WORK!? THOSE RETARDS THAT MADE NWN SHOULD HAVE MADE IT RUN ON ATI HARDWARE IN THE FIRST PLACE! OPTIMIZING FOR ONE COMPANY IS STUPID!!!

I can't wait until some ATi optimized games come out using proprietary ATi extensions so they don't work for the GayForce line of cards. Then I'm gonna bitch and say "nVidia makes bad drivers!".

I always hated 3D coders, they're such little b!tches. I know I'm stereotyping, but all the big hot shot ones (Carmack) deserve a BIG foot up their ass. There should be laws against paying for optimization on games, OR something along those lines.

Note: Humus is cool.

Let's try to watch the language...

[Bigus Dickus]

Originally posted by Chalnoth
You don't seem to understand. I meant to say that there is an ATI card out there somewhere, with some driver version, that should most certainly work with that code. In other words, it's only not working now because ATI's drivers in the past (Or with another card) were too lenient, apparently (coupled with sloppy coding, but if it works, why bother to troubleshoot it? But...apparently somebody didn't do enough testing...).

I do understand. And, aside from the validity of my previous comment, I can only ask "and you don't think this happens with nVidia drivers?" In fact, I remember reading some comments about "odd" things in nVidia drivers that were left "odd" out of fear that it would break compability with games that currently worked, and leave a negative impression with consumers.

I suppose they were correct. You bash ATi for having driver bugs, and still somehow find reason to complain when they fix those bugs. Well ****, it's a no win situation I guess. At least, as far as you're concerned. You'd bash ATi products even if they spit out gold coins once an hour. :rolleyes:

[Chalnoth]

Originally posted by Bigus Dickus
I do understand. And, aside from the validity of my previous comment, I can only ask "and you don't think this happens with nVidia drivers?" In fact, I remember reading some comments about "odd" things in nVidia drivers that were left "odd" out of fear that it would break compability with games that currently worked, and leave a negative impression with consumers.

Well, I don't. The only work-around for an nVidia card that I'm currently aware of is for the TNT in Vogel's OpenGL renderer for UT. If you can find another specific example, go ahead and post it. I'm not aware of any, so until one is put forth, it just seems obvious that nVidia's drivers are better. This is what every developer I've seen comment on the subject has said.

I suppose they were correct. You bash ATi for having driver bugs, and still somehow find reason to complain when they fix those bugs. Well ****, it's a no win situation I guess. At least, as far as you're concerned. You'd bash ATi products even if they spit out gold coins once an hour. :rolleyes:

Try: It shouldn't have been there in the first place. Just like the person who coded the ATI support in NWN was careless, so were the ones who didn't fully test the basic functionality of ATI's drivers, at some point or another, as it relates to this particular issue. In other words, ATI's driver team was every bit as careless as that NWN coder.

[AGP64]

If I look at the Catalyst sucks, Detonator rock discussion from a broader perspective I draw the conclusion that both are good, fast, stable and have some issues to resolve:

Det:

http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=4718&perpage=25&pagenumber=1

http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=2503

Catalyst:
http://www.rage3d.com/board/showthread.php?s=&threadid=33652267
http://www.rage3d.com/board/showthread.php?s=&threadid=33647282

Poll:

http://www.rage3d.com/board/showthread.php?s=&threadid=33651887

[Joe DeFuria]

Try: It shouldn't have been there in the first place.

No, really?

Now you are being critical because ATI drivers had bugs?

And yet nVidia has zillions of versions of its detonator drivers...I guess none of these fix bugs that were in earlier verisions. nVidia must have been really careless.

Come on, Chalnoth. Give it a rest already.

Bugs happen. You find them, you fix them. They happen in ATI drivers, they happen in nVidia drivers. nVidia has had it easy the past couple years because they have been the dev platform of choice due to the consumer popularity of its hardware.

Things are going to change over then next year or two...

[ATI LoVeR 9700]

Originally posted by Joe DeFuria
nVidia has had it easy the past couple years because they have been the dev platform of choice due to the consumer popularity of its hardware.


No they've had it easy because ever since the Geforce1, the core has been the same. The Geforce2 was an overclocked Geforce1 with some stupid enhancements. The Ultra and Pro were just SUPER overclocked.

The Geforce3 had pixel shaders and memory optimizations added, not anything really new, just additions.

The Geforce4 just had crazy high clock speeds, better memory optimizations, better Anti-Aliasing.

ATi on the other hand...

The Radeon is totally different from the Radeon 8500 and then the Radeon 9700 VERY VERY VERY different from the Radeon 8500. ATi has a whole NEW core every year. nVidia has one every 2 or 3 years. The NV30 is their new core on which for ages they will build upon...

[Chalnoth]

Originally posted by Joe DeFuria
Now you are being critical because ATI drivers had bugs?

Yes, because I still can't see the water in NWN on the 9700. ATI is half of the cause of that.

And yet nVidia has zillions of versions of its detonator drivers...I guess none of these fix bugs that were in earlier verisions. nVidia must have been really careless.

Um, of course there are bugs, but usually they revolve around global stability issues, things which are much harder to track down than simple rendering errors. Usually when somebody tries a new driver set from nVidia, there are complaints with things that are more out of the ordinary. With ATI's drivers, I'm seeing things broken that should never be broken in the first place (basic rendering errors).

And your argument that nVidia has been the development platform might hold if developers haven't stated that nVidia's drivers just plain better than ATI's.

[AGP64]

Originally posted by Chalnoth
And your argument that nVidia has been the development platform might hold if developers haven't stated that nVidia's drivers just plain better than ATI's.

Duh..:o

I read some weeks ago (do not have the link as I forgot to bookmark) a M$ product manager stating both the det's and the catalyst are on par. So depends who you ask.

Possibly "The way games are meant to play" sponsered developers state that NvIDIA drivers are 'just plain better'' :rolleyes:

[netape]

Originally posted by ATI LoVeR 9700
No they've had it easy because ever since the Geforce1, the core has been the same.

lol :D , you've got to be kidding, riiighht?

[-=DVS=-]

Nvidias Drivers sucked big time in Geforce 1 Age , Offcourse ATI wasn't competitive back then 3Dfx{uber company} was , but back to subject Hopefully when NV30 is released we woun't have to live throu crappy driver era like we did with GF1 :D :p

[jbirney]

I remember reading some comments about "odd" things in nVidia drivers that were left "odd" out of fear that it would break compability with games that currently worked, and leave a negative impression with consumers.

Hmm I remember in one of those 20+ pages post with good ol Derek where he said practrally the same thing.