• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

ATi Catalyst 3.2 Driver Released.

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Works very well for me, boosted my 3DMark2001se score by 700 points and fix my studder problem with BF1942

1.2Ghz T-bird > VIA KT266A > Radeon 7500 AIW 64MB > 512MB DDR Ram PC2100 > Windows XP Pro
 
BF1942 locks up with these drivers on my system :|

Anyone else getting issues in BF1942?

EDIT:
Nevermind, reinstalled it and it's fine now
 
Since I'm having some issues with my old 7200 le (and replacing it in a day or two when Fedex comes) I decided not much harm in trying these new drivers even with the old card and 98se. ATI installer crashed, which a number of people experience, but it did so after writing files to the .../ati/support/ directory, and manual installation of the .inf worked fine using device manager.

Without benchmarking I can say that open gl performance is a bit smoother in a couple of older games I'm playing right now. If I were planning on using this setup for a long time though I'd be concerned about the number of issues being posted at rage3d.
 
After the usual "installation from hell" with Win98SE, these drivers are working nicely for me. 2D quality is greatly improved, text is very sharp and stable.

My 3D2001 scores dropped by about 300 points, but overall quality during all tests was improved. When O/C'd, I noticed that the usual "tearing" I get during these tests has dropped by about 75%! (My 8500LE has ridiculously slow memory!)
 
After extensive testing I've determined that the Catalyst 3.2s are pretty much the same as the Catalyst 3.1s in terms of performance and outstanding issues except the mouse response is much better in NOLF. It's still a bit laggy though.

Sac is D3D, can't force 24bit Z unfortunately.
For Direct3D mode if you run in 32 bit coloure the drivers will automatically force a 24 bit/32 bit Z buffer. Also check the game's settings as well.

I suppose it is possible it is a hardware bug that ATi can't fix,
I highly doubt it. Sometimes it does take ATi a while to respond to reported issues.

sim-city 4 is broke? what did they do to it?
I believe that latest game patch has fixed the issues on Radeon cards.
 
3.2 kicked me out of games, gave me a blue screen, and rebooted my machine.

Switched back to 3.1 and all is good.
 
drivers worked fine for me. even let me overclock my 9700 a little more (now at 337.5/337.5).
UT2K3 seems to run faster (~ 35fps with AA off, 16X performance AF, 1280x1024). 3dmark2K3
gives me a little over 4000 with the 16X AF on.

I'm pretty thrilled with the drivers. If the 3.3s are as good as people are saying, I'll be a happy camper.


 
Fixed a stuttering problem for me in Freelancer whenever I moused over something that would pop up text.
 
"For Direct3D mode if you run in 32 bit coloure the drivers will automatically force a 24 bit/32 bit Z buffer. Also check the game's settings as well.

I've checked the game settings numerous times, tried different drivers, downloaded RageTweaker and had no luck with anything, including a mixture of all of the things I've mentioned.

I highly doubt it. Sometimes it does take ATi a while to respond to reported issues.

Roughly six months now. Are their driver teams really that bad? If it isn't a hardware bug, then ATi still can't carry nVidia's jock strap when it comes to drivers.
 
Originally posted by: PClark99
3.2 kicked me out of games, gave me a blue screen, and rebooted my machine.

Switched back to 3.1 and all is good.

what games did it boot you from???

do you need smartgart to disable fast-writes on a gigabyte 8inxp??? i dont want to install it if i can disable them from the bios
 
Roughly six months now. Are their driver teams really that bad?
Well it's true that ATi aren't perfect but I'd imagine that the game you listed isn't very popular and thus hasn't got a high priority.

If it isn't a hardware bug, then ATi still can't carry nVidia's jock strap when it comes to drivers.
I couldn't run GLQuake higher than 1280 x 960 x 16 for around 6 months on my nVidia boards. Also Serious Sam 1 & 2 both had 16 bit Z issues on my Ti4600 which haven't yet been resolved.

The devs blame nVidia and nVidia blames the devs. I don't know who's at fault but I do know that the problem does not exist on any Radeon board I've tried. And don't even get me started on the DXT1/S3TC issue, a problem which still exists in GF4 boards despite nVidia's hardware dithering to mask the problem.

The point is that no vendor is perfect but I'd certainly agree that nVidia still has a slight edge when it comes to drivers.
 
Well it's true that ATi aren't perfect but I'd imagine that the game you listed isn't very popular and thus hasn't got a high priority.

It won numerous GOTY awards and is in the top 40 PC games of all time on GR. http://www.gamerankings.com/htmlpages4/3730.asp

I couldn't run GLQuake higher than 1280 x 960 x 16 for around 6 months on my nVidia boards. Also Serious Sam 1 & 2 both had 16 bit Z issues on my Ti4600 which haven't yet been resolved.

Were you using a ref fix utility? I know that broke 16x12 for me on quite a few games, although uninstalling it would allow me to play at the settings I wanted(including Quake). As far as Z errors, I don't run 16bit in anything that I can recall so I don't know about that(although Sac defaults to 16bit and doesn't suffer from the same level of Z fighting on nV hardware that ATi's boards to running 32bit).
 
It won numerous GOTY awards and is in the top 40 PC games of all time on GR. http://www.gamerankings.com/htmlpages4/3730.asp
In that case you might try posting at Rage3D to see how far on ATi's priority list this is. Sireric and Catalyst Maker are very good in that respect.

Were you using a ref fix utility?
Excellent guess and yes, I was. However if I took it off GLQuake still wouldn't work and the only solution was to go back to very old drivers. Of course this was unacceptable. Then just before I upgraded to my Radeon 9700 Pro from my Ti4600 nVidia released new drivers which finally fixed the issue.

None of the Radeon boards I tried had this problem, with or without the refresh fix, regardless of the drivers I used.

As far as Z errors, I don't run 16bit in anything that I can recall so I don't know about that
Neither do I, this is a problem with GF4 based cards running Serious Sam 1 & 2 in 32 bit mode. The game requests the same level of bit depth as the colour level and if the board doesn't support that level then the drivers should return the next highest precision. The Radeon does 32 bit just fine, the GF3 doesn't but correctly reports 24 bits which is also OK but the Ti4600 reports a 16 bit Z-buffer. As soon as I dropped in my Ti4600 into my system after having a Ti500 I was greeted to sawblades on polygonal joins all over the place.

FYI there is a console fix for it by manually forcing the game to a 24 bit Z-buffer. This is a known issue and Croteam says it's nVidia's fault while nVidia blames the application. I don't know who as fault but I do know that my Radeon works perfectly.
 
In that case you might try posting at Rage3D to see how far on ATi's priority list this is. Sireric and Catalyst Maker are very good in that respect.

I've talked to one of the ATi driver team members at B3D about the situation. I know they are aware of it, but I haven't seen any indication that they ever intend to fix it.

The Radeon does 32 bit just fine

The R300 doesn't support 32bit Z, it only supports 24bit. If the Rad is running the game OK, then obviously the game can utilize a 24bit Z without forcing it.
 
I've talked to one of the ATi driver team members at B3D about the situation. I know they are aware of it, but I haven't seen any indication that they ever intend to fix it.
In that case I wonder if it's a W-buffer issue instead of a Z-buffer issue ala Sim City 4? It may need a patch from the game.

The R300 doesn't support 32bit Z, it only supports 24bit.
All of the Radeons I've tried worked though I specifically didn't check the game for bit depth with the 9700 Pro. Maybe it was 24 bit, which is still just fine.

If the Rad is running the game OK, then obviously the game can utilize a 24bit Z without forcing it.
Yes it can, it'll use the highest value reported by the drivers. The issue is the GF4 reporting 16 bit all the time.
 
Back
Top