No Windows 2000 support for the ATi Rage Fury MAXX

[Gurm]

Dells blow. Although, the GX1's we have are running Win2k with no MAJOR problems, albeit with some features turned off (power management, and the Zip drives are flaky... grr...)

Win2k rules.

ATI still can't make a driver that doesn't have serious 2D corruption in Win98! Nevermind Win2k! Come on, ATI... get with the program. You can't just make a card and then not make good drivers EVER!

- Gurm

------------------
Listen up, you primitive screwheads! See this? This is my BOOMSTICK! Etc. etc.

[The PIT]

ATI + Win2k = Recycle Bin.

[Arsa]

W2K works nice for me (even on K7V ;-)
We had some problems installing it on Latitude (had to upgrade BIOS), but nothing else... Oh, and IBM Netfinity 5000 works great too.

[Nuno]

So I assume I can assemble PC´s better than Dell...
Win2k runs like a charm in my Athlon. I still have win98 in dual boot, but sincerely I can´t remember the last time I booted that OS.

Joel, something must be wrong either with the Dell PC´s (imcompatible hardware?) or the way people are trying to install win2k in those machines. Win2k is very stable and is much more compatible and easy to use than Microsoft itself preaches.

[kjliew]

It seems that Win2k is very particular about display hardware implementation. Non-standard multi-chip implementation may not pleased Microsoft. This has happened on Voodoo2, and now ATI Fury MAXX. And, it's kind of doubt if 3Dfx will ever come up with Win2k drivers for multi-chip Voodoo5.

If I remember correctly, Matrox next solution for gamers, G800, is also a multi-chip solution. Let's wait and see.

[C=64]

I made a flippant remark that was meant to be taking with some HUMOR and to vent a tad about my disappointment with Matrox's Win2k development. I felt it would be kinda humorous to use the ATI message as a basis for such, considering LAMFDTK had so graciously made it available.

My complaint, Maggi, is squarely aimed at issues that should not exist in the G400 Win2k drivers. Further, I do NOT consider my postings to be "nonsense" or intending to "confuse/mislead other people". I thoroughly resent these implications.

If you do indeed require me to further expound upon my statement to satisfy your judgment as to my fitness to also post on this messaging board Maggie, I guess I will.

First, the lack of SMP compatibility is UNFORGIVABLE. Period.

2nd, the crippled dual head support is a rather sore point for me. Some of you may point out that the dual head support works just as well as it did on NT. Fine. But I will point out that Matrox's advertising classifies Win98's dual head as a complete implementation, and that Win2k, UPON it's release, was to be privy to the same dual-head lovin', in stark contrast to the incomplete NT support. I'm afraid these aren't the exact terms that Matrox uses in their advertisements, but I think they give the basic idea. Not that I use the present tense. This is intentional.

Now we're being fed the line that the first service pack will correct the issue. Some of you may say that this is an acceptable compromise. I, unfortunately, do not share such an opinion on the subject. If a service pack is deemed necessary, this should have been noticed earlier. Much earlier. It should NOT have been left until AFTER the release of the FINAL product became available on mass to the public. I, being the poor-spirited pessimist that I am, feel like this is some sort of stall tactic, to sate our minds while our bellies remain empty, giving Matrox more time to work on the driver.

Now that this holy grail of a service pack is being beta'd, I'm awaiting, with one hand clutching my chest and the other my wallet, for word from Matrox in either direction, so at least I can feel comfortable with whatever decision I am forced to make.

I purchased my G400 Max before Win2k's full release, putting full faith in Matrox's claims. While I'm glad that Matrox has not abandoned driver development on the platform for my chipset like ATI has done, I still don't feel I should be cheering for Matrox just yet. The only thing that is appropriate is jeering ATI for their lack of foresight, commitment, and respect for us, the consumer relegated to depending upon a company's whims and fancies.

To close, I'm sorry my bit of lax humor at the expense of ATI was just too much for your delicate sensibilities, Maggi. I’ll try to make sure my future statements are bland, uninsightful, and full of nothing but praise for everything Matrox.


C=64

p.s. All this sentiment of blaming canada brings to mind a jolly little melody. How about a nice rousing rendition of one of everyone's favorite South Park tunes?

Sorry for such a long post. Too much caffeine, not enough code.

[This message has been edited by C=64 (edited 12 May 2000).]

[This message has been edited by C=64 (edited 12 May 2000).]

[Rags]

PaulS,

Now, that is an entirely different way to look at things. You think that because they have released 3 more vid cards in the last year, that excuses them from properly supporting the TNT2? That is a HUGE turn-around from your many other posts that contradict this new-found belief that poor support for a relatively young product is fine, just so long as the new flagship recieves full support. Uhh...yeah, forced obsolescence is real good, now isn't it. I will stick with my Matrox card, thank you very much.

Rags

[paulcs]

Well, I test a GeForce DDR with Windows 2000 and Quake3 over the weekend. The results were pretty much the same as they were with the TNT2U. The 5.1X W2K Detonators don't seem to support SMP at all. I tried three different sets, and none of them worked.

The 3.84 Detonators worked, but I found them to be less stable with the GeForce than the TNT2 Ultra. The bugs remained the same: if SMP was enabled upon leaving the game and rebooting, the game would not load and would lock up at ID's noisy introduction screen. SMP had to be disabled directly in the config file or you could just forget about getting into the game.

Lockups were very common, although not during play. They typically occured randomly and at the startup screen, the main menu, while a map was loading, or at the console. They seemed to be more frequent at 32-bit and higher resolutions.

Image quality was pretty bad. The "SMP advantage" was most pronounced at lower resolutions and color depths, as you would expect. It was often not a factor at high resolutions and 32-bit, although the map could play a role.

NVidia recently released a new Quake3 map, and strangely enough, it seems designed to make their own boards choke. (I'm not living under any illusions here. It seems designed to make every other board choke as well.) At high resolutions, the two CPU's were still a factor when benchmarking with this map, and when frame rates are this slow, an additional 3 or 4 FPS really helps. For the most part, however, when the going got tough, the second CPU stopped being a factor. Framerates above 800x600, 32-bit, were identical or nearly identical running the Quaver timedemo with all the settings on high.

Given the GeForce2's memory bandwidth limitations, I suspect the 32-bit scores will be similar. (I'll know in about a week.) I also expect the SMP support to be buggy as well. This appears to be a driver issue. I don't know what kind of crap W2K programmers are going through, but this is just not working as well as I would have hoped. To my mind, nVidia appears to be moving backwards as far as SMP gaming support goes.

Paul
paulcs@flashcom.net

[This message has been edited by paulcs (edited 16 May 2000).]

[Pace]

Although I agree with you in part C=64, this remark seemed to continue the Matrox flame:

Now we're being fed the line that the first service pack will correct the issue. Some of you may say that this is an acceptable compromise. I, unfortunately, do not share such an opinion on the subject. If a service pack is deemed necessary, this should have been noticed earlier. Much earlier. It should NOT have been left until AFTER the release of the FINAL product became available on mass to the public.

Microsoft released Win2000, Microsoft are releasing SP1, Microsoft released the bugs, Microsoft knew about the bugs. Although Matrox may have to account for some of the blame, I feel that this problem seems largely Microsoft's. And that this was to give Matrox time to produce a driver - no, I I think not.

Paul.

[PaulS]

Rags,

I'm not exactly stunned with my G200's performance on W2k, but I accept the fact that this card is a generation behind the G400. What chance do I have of getting Matrox to release top notch drivers for this card when the failed to do so with NT, and are failing to do so with the G400 on W2k?

I was reading the review on the RT2000, must break the heart of the engineers that designed that card to read how great it is...except don't try to use it with NT.

Personally, I wish everyone would stop working on new cards and just concentrate on fully exploiting the potential of the existing ones (yeah I know, no money in that). So far we are just trying to get Matrox to fix the bugs, I wonder how long it will be until the drivers are optimised to scale well with SMP systems. The G400 scales well with faster processors - the same should be true for multiprocessor systems.

Professional cards optimised for toy operating systems. Pretty much sums up what Matrox has been doing for the last couple of years.

Paul

[Rags]

Hi PaulS,

I agree with you, I wish cards would be optimized fully before moving on to the next gen. too.

Rags

[C=64]

Although I agree with you in part C=64, this remark seemed to continue the Matrox flame

I'll admit, I should have toned that down a tad. I made a posting boo-boo; I posted while I was still upset. My bad.

However, while I don't stand by the tone, I do stand by the message of that remark. Matrox could have told us a lot sooner about the current difficulties with Win2k. Unless they announced it before RC 2 of Win2k, it wouldn't have made a difference on my part, but I am sure there are plenty of others who decided on the purchase afterwards. But the last time I checked (which was fairly recently), Matrox's posted specs and information are still claiming full dual head support. And, quite frankly, this is unacceptable. If the feature doesn't work yet, well then it doesn't work yet. But do not lie to me and say that it does.

I have a hard time seeing this as an OS bug. Call me a pessimist, but I don't think this forth-coming service pack is going to correct anything. My reasoning rest partially on ATI's announcement. If this problem is solely "issues with enumerating the chip ID's" under W2k, and a simple OS update would correct it, ATI sure did a good job of shooting itself in the foot. I'll be overjoyed if I'm wrong, but at least I won't be crestfallen if I'm not.

My paranoia about Matrox's driver development could be easily sated. One thing that upsets me about the Matrox driver development cycle is that it's pretty closed. Matrox won't even give a time frame of when to expect new driver. For drivers from Nvidia, though, I see information all over the place about what's going on with them, what their current difficulties are, work-arounds, release dates, set backs, etc. And all with what appears to be the companies blessing. If only we could see something similar from Matrox.


Rags and Pauls, do you remember SciTech's stuff from back in the good ol' DOS days? If a company wasn't willing to spend time on their old products, at least SciTech as willing to pick up the slack. Maybe if our modern batch of companies were willing to release the source to their drivers, a modern day SciTech (or heck, SciTech themselves ) would commercially exploit the situation, and at least give us SOME ease of mind.


I hope I used a more reasonable and collected tone this time around.

C=64

p.s. One thing I forgot to add to my previous, fairly hostile message. Maggi, I really don't think PaulS is in a position for me to confuse/mislead. I've followed his postings in other threads, and I've found him to be both knowledgable and frank in his statements.

[PaulS]

C=64, to some extent that has already happened. The open source Linux community did a pretty good job on their OpenGL drivers without access to the most important chip code.

I'm not sure why companies are so reluctant to open source their drivers, maybe they're afraid it would help competitors to reverse engineer some of their technologies?

Did you notice Matrox's little article praising the virtues of open source? Now if only...

Paul