Announcement

**neepheid** · 23 May 2000, 01:58

Hi there

I'll await the outcome of this thread with much interest. I haven't got a Matrox driver later than 4.51 to work successfully with my setup.

While I accept it could be an issue with my machine, it's happening a little too often to SS7 users to be coincidence.

Setup

AOpen AX59pro SS7 motherboard
AMD K6-III 450(500)Mhz processor
128Mb SDRAM (1x128)
Matrox Millennium G200 8Mb AGP
Diamond Monster 3D II 12Mb (Voodoo 2)
Creative PCI128 soundcard
Diamond Supraexpress 56i PCI modem
3Com 3C509 ISA 10M ethernet card
8Gb + 6Gb harddrives
Sony 32x CDROM
Mitsumi 8x4x2 rewriter

Windows 98
Latest VIA drivers
DX 7a

I wonder if we can get this ironed out.

neepheid (R)

If you have to ask.... oh well
It's a Scottish expression for a person with no common sense. Literally: turnip head.

**[Bozo]** · 23 May 2000, 13:49

Interesting... I have had exactly the same problems. That is, blue screens in 3D with 5.30, 5.52 and 5.55 drivers.

I have tried several versions of the VIA AGP driver and several version of BIOS for the motherboard.

The thing is that I also have the AOpen AX59Pro motherboard. This is the version with the first MVP3 chipset (CD).

I think matrox is using strange bus mastering features because if I turn off bus mastering, I get no blue screens in direct3d (as far as I know, OpenGL reqs. bus mastering) although much slower perf.

I have had a GeForce card in the same machine, and it worked perfectly with AGP2x and bus mastering so these problems cannot be explained by the MVP3 chipset alone.

It's also very strange that matrox introduces this kind of bug in certain versions of their drivers, without fixing them in subsequent releases.

I have seen many reports of errors from socket7 users, this isolated issue might have something to do with AX59Pro, but there is so many reports that matrox obviously doesn't care about socket7 users.

**neepheid** · 24 May 2000, 02:24

Hi there

I thought I'd give the latest (5.52) drivers a go again. I did a clean install of Windows 98 (setup -p j for ACPI stuff) and then installed VIA drivers (4.20). Finally, I installed Matrox drivers. Oh no.

Everything seemed to go OK when I came back into Windows. Except the mouse pointer would occasionally jump to other locations on the screen of its own accord. I know some people with Logitech scrolly mice were complaining of issues, I have a Genius Netmouse Pro. I had not yet installed the latest drivers for the mouse, so I did so and that seemed to cure that problem. But the killer was when I tried to use OpenGL screensavers.

When I tried to set a screensaver (eg. 3D Pipes), as soon as I selected it there was a freeze of around 7 (yes seven) seconds before the preview came up. Also the mouse pointer started blinking during this process. When I hit the preview button, I had to wait a further 2-3 seconds before the screensaver kicked in, during which time you could not get back to Windows.

Oh, and one BSOD.

I don't know. I uninstalled these drivers and put back in 4.51. What a surprise, no problems. But this is unacceptable, because these drivers don't know about DirectX 7 and DX does not realise that I am using an AGP card.

This sucks.

------------------
neepheid (R)

If you have to ask.... oh well
It's a Scottish expression for a person with no common sense. Literally: turnip head.

**[Bozo]** · 24 May 2000, 03:42

It's almost a known fact now. Several versions of the 5.x drivers for G200 (mabye G400 too?) has severe problems with MVP3 chipsets and 3D accelerated modes. I'm almost 100% sure this has to do with the way matrox uses bus mastering functions.

As for the "mouse stuttering" this is a powerdesk issue. Turn it off, and there is no mouse stuttering.

The only thing left to investigate is if early revisions of MVP3 chipset has more problems than later. As far as I know the early revision has letters "CD" printed on the chip. (many AX59Pro boards has the early revision)

People claiming that the problems only has to do with the MVP3 chipset are mistaken. A properly configured MVP3 system works great with many brands of graphics cards (like the GeForce). These problems has only occured with some versions of the PD 5.x drivers. Older drivers (4.x) never had these problems.

The only known issue with the MVP3 chipset is signal-noise making faster transfer modes unstable (i.e AGP2x). However some graphics cards cope with this issue better that others (the G200 does not cope with it at all)

And for all those who has problems with properly installing VIA drivers... use the built in (w98) util System Information to determine which drivers are in use. PCIList is also another good program to examine AGP functions.

So... use PD 5.41 driver because these are stable with MVP3. I have only had issues with 2D modes with this driver, and this can be solved by turning bus mastering off for 2D modes *ONLY*. (yes there is a seperate registry key for this and no it does not inpact on performance)

**DentyCracker** · 24 May 2000, 10:12

DIME vs DMA. DIME is less tolerant of noise than DMA but is theoretically better. GX00 uses DIME hence .....
If you want to try to get the card working properly the best bet is to follow Rags' instructions. He has posted them several times on the fori. Do a search.

**[Bozo]** · 24 May 2000, 11:12

Even if DMA mode would solve noise problems, this would only enable us to run the G200 in AGP2x mode on MVP3 mainboards (most MVP3 users knows that running the card to AGP1x is the only option).

This would of course be nice, but will not solve incompabilities with the above mentioned drivers. Note this problem is not instability, rather an incompabiliy between the VIA drivers and certain PD 5.xx driver versions. As I mentioned, I suspect they way matrox uses bus mastering in these driver versions to be the reason for this.

However it would be intressting to read about these "Rags' instructions"... if I could find them that is.

**Angel2167882** · 24 May 2000, 22:26

Hi again Folks!

I forgot to mention... my AX59Pro have the "CE" MVP3 rev. chipset.

The first AX59Pro I bought (in summer 98) rev. #1.1 (I think) got the "CD" rev. After heard of the problem with the "CD" and i740 video card, I immediatly call AOpen for a m/b replacement (my reseller would'nt do it).

So, AOpen said ok (I said my reseller was no more), but it cost me 75 CND$ for the custom fee (Where is the "free exchange") + shipping...

So, now I got an AOpen AX59Pro 1MB L2 (my first was a 512K L2) rev. 1.5 with the "CE" rev. chipset.

And by the way... thanks you all for the repost... keep in touch if you have news about the Powerdesk 5.55... I will try the procedure leave by a Matrox techician replying to me at the Matrox forum... look below:

***PASTE BEGIN
Try our uninstaller and remove our drivers. Then go directly in safe mode, and search for PD*.* Delete pdesk, pdkernel.dll, pdeskres.dll, pdtools.dll and pdesk.lgc. Then go in your device manager and remove all display adapters and monitors listed and reboot. Then reinstall the 5.55 drivers to see if the problem persists. Also if you have the 4-in-1 driver 4.20 installed, you should also install the individual AGP Driver 4.02. Please make sure that there is nothing installed in the first PCI slot next to the AGP slot.

------------------
Dennis
Matrox Graphics
Technical Support Rep
PASTE END***

and ironicaly, there is an answer from another Matrox technician saying the opposite... (see the story at http://forum.matrox.com/mgaforum/For...ML/001132.html )

***PASTE BEGIN
You will need to :
- Uninstall your drivers with the uninstaller
- Boot directly in safe mode
- Make a search for pd*.* and delete our 3 remaining files
- Go in your device manager and delete all display adpaters and monitors
- reboot and go in your system BIOS
- Be sure that USB is enabled!! (this should have been done before you installed windows according to VIA)
- Boot in normal mode
- Upon detection of the card install drivers for standard VGA adapter from windows list only. Do not install Matrox drivers at this point
- Install the latest 4 in 1 patch from VIA
- Reboot
- Install VIAINF file
- Reboot
- Install our 5.41 drivers (rather than 5.52)

------------------
Alex
Matrox Graphics
Technical Support Supervisor (Graphics)
PASTE END***

So, Matrox tech are theirselves confused over there one saying try the new 5.55 and another saying NO NO NO... DON'T use the 5.52, use the 5.41... what the fuc*...

See ya all

[This message has been edited by Angel2167882 (edited 25 May 2000).]

**[Bozo]** · 25 May 2000, 13:35

Oh.. so it does not seem to be any difference with revisions of the MVP3 chipset. There still might be some strange thing going on with the AOpen AX59Pro, but I do belive that users with other motherboard brands (with MVP3 chipset) share the same problem.

That only leaves one thing to conclude... 5.30, 5.52 and 5.55 drivers are incompatible with the MVP3 chipset *OR* the VIA AGP miniport driver. (I would like to test this on a ALI chipset mobo.. they probably doesn't have the same incompability..) In either way... this is a software bug and can be fixed (why in hell would 5.41 work otherwise?!)..

Other gfxcard vendors have worked with VIA to solve such problems.. It would seem as if matrox is reluctant to do this...

Announcement

Why all new drivers "suck" since the 5.25 (for G200)

Why all new drivers "suck" since the 5.25 (for G200)

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment