It´s only happening here, or all win2k 5.3x drivers disable AGP?

[Gurm]

Well, it's not happening here. AGP is enabled (at least according to the DirectX tools it is), with the VIA 4-in-1 4.28a and the MS AGP patch and my MSI 694D (dual P3).

Is this really an Athlon vs. P3 problem? Hmm... weird!

- Gurm

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

[Nuno]

I would bet more on a VIA vs Rest of the World problem. VIA takes too much shortcuts on their hardware/drivers. Matrox win2k drivers are to blame, surely, but I could bet they didn´t nothing than expose another VIA AGP implementation problem.

It would be interesting if someone with an old 750/Athlon classic or with a new 760/DDR system could report how 5.3x drivers are doing in their systems.

You can bet this is the last time I buy a VIA board. I had nothing with VIA boards but trouble. The older gigabyte AMD 750 board I got was FLAWLESS with everything I could throw at it. Now with this POS (and the A7V is an expensive POS) in the last month I spent more time troubleshooting the system than using it.

[This message has been edited by Nuno (edited 28 January 2001).]

[DGhost]

I admit it would be very interesting to see what the results of it on an AMD chipset are like. But, as i have pointed out already, i run a MSI 694D which is based on the Pro133A chipset, the 'sister' of the KT133, and the 5.33's work excelent. acctually, they work better than any so far. VIA being VIA and cutting corners has made sure that the drivers between the two chips are compatable, even to the point that they use the same drivers. I know that doesn't mean alot, as the G400 and G450 use the same drivers as well, but they should be fairly compatable, if not identical.

Also, the fact that the A7V's were a pain in the bum is the major reason that we are selling EPoX boards at work - they are very good quality, stable boards.

[Rob M.]

Another w/disabled agp texturing here

I'm a7v also, win2k sp1, via 4in1 4.27, matrox 5.33 on my g400 max, AMD registry patch, microsoft 'pre-sp2' via exe patch. I just did a clean install of my system, I don't have any traces of any drivers other than the latest around...

[Ashley]

<font face="Verdana, Arial, Helvetica" size="2">I would bet more on a VIA vs Rest of the World problem. VIA takes too much shortcuts on their hardware/drivers.</font>

I have the same problem with the 5.33 driver on a *BX* mainboard...

[Ashley]

In the case of my BX,

Mga.AGPFlags 00 00 00 00
Mga.AGPFlags 04 01 00 01 (the driver default if I use a G450)

both leave AGP texturing disabled.

Mga.AGPFlags 01 01 00 01 (the driver default if I switch to a G400)

enables AGP texturing.

[Rags]

Mine says 04 01 00 01, and AGP is enabled.

This is under win2k using the 5.3 drivers.

Rags

[DGhost]

my G400MAX on 694D says 04 00 00 01, AGP is enabled @ AGP 4x w/ 5.33.

could it be (agp speed) (autodetect) (?) (SideBand Addressing??). Mine is currently forced to AGP 4x and sideband addressing is on.

HMMMMM

[dZeus]

as I said it's working properly on my G400.. here's my MGA.AGPFlags value:
01 01 00 01

It's running at auto-detect speed, and goes to 1x AGP on my overclocked AGP bus (89MHz). Not that I care, since I get a 3DMark score of 3439 with the PD5.33 drivers in Win2k , and OpenGL performance is also top-notch!

[Olivier]

I reinstall the 5.33 on the 5.20 and again the same problem; bad graphical performances and agp texturing disabled..
Back to 5.20... again...

[Ashley]

In my case, the driver defaulted AGPFlags to "04" (first byte is transfer rate).

Since I was using a BX mainboard, that default is just invalid - hence, AGP texturing was disabled. Changing the first byte to "01" or "02" fixed the problem.

Just guessing but if your default value for the first AGPflags byte is "04" and AGP texturing is working just fine, it may be because you have a 4x-enabled mainboard.

[Nuno]

Yes I also figured out what that "4" would be. But anything I enter in the registry can´t get me AGP with 5.33. I tried several combinations that you people reported to work and it´s a no go...

[Nuno]

Just a quick update: Dx8a doesn´t fix the problem.

[Nuno]

Ok Ok, another update
Matrox TS guys have been able to reproduce the problem and (hopefully) we may expect updated drivers with a fix.
Cool