View Full Version : (Epox) MVP3 boards + G200 + PD5.30: Disable busmastering will solve crashes

Randy Simons
12th October 1999, 07:50
Dunno if this is already known to you guys, but in a reply to my mail to matrox about crashes with MVP3 Epox boards and G200 with PD 5.30 they gave me a tip which I had forgotten to test before: disabling busmasting.

So, disable the busmaster option for your G200 and things probably will run fine again, without crashes with PD5.30.

This will result in a performance hit, but at least it works again http://forums.murc.ws/ubb/smile.gif I hope Matrox will fix things soon, so we can use busmastering again.

Hmm, just found out that OpenGL won't go in hardware-mode.. XTom3D switches to the Microsoft GDI Generic OpenGL drivers (software)... anybody else with a solution for this?

Randy Simons
Digital Dreams Software http://www.dgdr.com/

[This message has been edited by Randy Simons (edited 12 October 1999).]

12th October 1999, 08:21
OpenGL hardware acceleration will NOT work w/o bus mastering. Disabling bus mastering should never be considered a "solution", no matter what a manufacturer says.


Randy Simons
12th October 1999, 08:35
Too bad...

Hm.. got reply from Matrox again (within an hour... WOW MATROX, this is not what I'm used to! :-)

The answer was less positive:

Dear Sir,
This problem is not our fault but the fault of all AMD based motherboards,
since the chipset makers are not following AMD's instructions. This
weakness is way there are even driver files in the first place.

Hmm, I recall my mainboards was one of the recommended mainboards by AMD... Weird stuff.

Well, rules or no rules, it used to work before, so why can't it work now?

[This message has been edited by Randy Simons (edited 12 October 1999).]

12th October 1999, 09:33
I suppose Matrox would say the solution is to buy a G400...? Or is it??

I have a MAX on its way to me from the Meadowlands UPS depo, should be here today http://forums.murc.ws/ubb/biggrin.gif (I hope... Thanks, Jord!!! Thanks, Pauly!).. does this mean it won't work (right) till I get my P-II/BX board set up (in the near but unspecified future)?

...well, at least I get to show off my new .sig http://forums.murc.ws/ubb/wink.gif....


All you people, can't you see, can't you see
How his love's affecting my reality
Everytime I'm down, he can make it right
And that makes him larger than life.
...not to mention being 6'5", eh Jord...?;)
What? System specs? Oh, yes, right...well, at the moment, it's: Soyo 5EMA, K6-2 333, 256MB PC100 SDRAM, Maxtor 7GB HDD, LS-120, Promise Ultra33 EIDE controller, Diamond Monster Sound MX300, HP 8100i CD-RW, Hi-Val (Nakamichi) 16X 5-disk CD changer, Win98 +SR1...oh yeah, and a Millennium G200 8MB SGRAM.

12th October 1999, 09:45
There is a less drastic measure than turning off busmastering for your G200. First try disabling DMA support for your CD-ROM or DVD. This worked for me and let me keep busmastering turned on for my G200. It slows access to your CD-ROM, but once accessed it runs just about as fast.

A better solution is to install the newest VIA 4-in-1 drivers. The new VIADMATOOL it installs lets you turn on UDMA support for your hard drive and CD-ROM separately. But, the new drivers have finally let me keep both turned on and keep busmastering for the G200 on as well. You might still have some stability problems, but if you do, turn off DMA on the CD-ROM first using the VIADMATOOL.


Randy Simons
12th October 1999, 10:43
Hmm, "desperate" as I am, I've tried disabeling the busmastering for my IDE device. It didn't work, as I expected, since I have both a PCI SCSI card (adaptec 2940UW) and PCI NIC (3COM 905B) which are both busmastering by definition. (thus can't be turned off)

I believe Matrox is talking crap: I've read no complains about G400's not working on Epox MVP3 boards. Or does nobody has such a setup? http://forums.murc.ws/ubb/smile.gif

There's only one solution: matrox, fix it.

12th October 1999, 10:51
Someone told me that it depends on what rev your MVP3 chip is. The later ones ("C" and "D" I believe) work better, but it's still a toss-up whether you get AGP 2X or not.


People call me a computer god; I remind them that I am merely a minor deity...

Randy Simons
12th October 1999, 11:13
There are 2 revisions on the board which are sold: CD and CE. Rev CD had major problems with Intel's i740 chip-based cards, and probably others too. But I've had no probs before with my CD revision board and the G200. AGP 2X never worked though.
Now I have the same board, but newer revision and I assume (never checked actually) I have the CE revision of the MVP3 chipset. Still no AGP 2X, but I find that much less important than something like busmastering.

12th October 1999, 13:29
I've been using an Epox MVP3G-M (rev 1.0) mobo and had absolutely -no- compatibility problems with anything until PD5.30
I have a G200, and was able to get agp2x mode and busmastering on and everything previously, but 5.30 crashed on any D3D or OpenGL games. Busmastering is not an option/solution in my mind because its always worked before, and I don't take well to dropping system preformance just to get something to work (when it should work w/o any bios hacking).
So.. here is a simple solution for all the PD5.30 problems and G200s... Switch back to PD5.25 like I did, and just wait for the next PD release. http://forums.murc.ws/ubb/smile.gif

12th October 1999, 14:52
Here's another "fix" I had to make that I forgot to mention, but it might help. VIA chipsets apparently only busmaster the two PCI slots closest to the AGP slot. I could only get my NIC to work in the slot closest to the AGP slot on my FIC 503+. Move your NIC and SCSI card to these two slots if you haven't already.

I also noticed that ANT posted a note on the MURC news that AGP 2x only has a chance to work on VIA boards if these two slots are only using busmastered devices. This is another reason to try moving cards.


12th October 1999, 17:52
I have an efa viking III mobo AMD k6-2 350mhz. g200 8megs sgram and have had no major problems with matrox drivers,even back in beta days,until now,with 5.30 ie.3d=bsod.
Went back to 5.25 and hope matrox comes up with a solution in their next drivers.
Seems to be too many people with problems.
All the boards having problems are AMD approved!So I don't buy matrox's line,think it's a case of stonewalling until they figure out what's up or down!

12th October 1999, 18:13
I buy Matrox's line, SS7 boards are some of the worst built boards out there, even if they are on AMD's recommended list. The approved boards are only tested for compatibility with AMD processors. EVERY full AGP texturing card is at a disadvantage on a SS7. Why do you think that DMA is a tough thing to achieve with a SS7? Why is it that SS7 boards have a tough time with NIC's, with new SCSI cards???? It is because the boards are TOO NOISY!!!


12th October 1999, 20:06
I haven't tried the 5.30 drivers yet. But the G400 on my Epox MVP-3G board has been the most reliable set up yet. Not that I had any problems with my old g200. Rags, there are good super 7 boards out there. They might take a little more care and feeding sometimes than Intel. And I have seen some prety bad BX boards in my time. (Dell's got some nice I820 boards to sell you cheap!!!).

Joel, any comments???

Epox MVP3G-M, K6/2-450, 256 meg PC100, G400 SGRAM 32 meg single, Diamond
Monster 3DII 8meg SLI, Ensoniq Audio PCI, WD 6.4 HD, Toshiba 32x, Toshiba 12x SCSI,
Diamond Fireport 20, Yamaha 4x6 CD writer SCSI, Generic NEC2000 network card,
Viewsonic E771 monitor Win95 OSR 2.1.

12th October 1999, 20:23

I have owned/built many SS7 systems, and still own one. There is no doubt that the vast majority of the SS7 boards are no where near what they could be. I believe that the mfrs. out there could build a good SS7 solution, but there really is not much incentive to do so. Even the best current SS7 boards are not living up to the potential of the platform. My G400 runs fine with PD 5.3 on my 503+ in 2X DiME mode. My G200 runs fine in 1X mode with the current drivers, and see no issues with PD 5.3.


12th October 1999, 20:51
A friend of mine went through 3 SS7 motherboards (FIC, EpoX, ASUS) before he finally took my advice and bought a Tyan.

Problem solved.

This machine is a S1590S and have had no UDMA/DMA problems, my NIC works fine, my G200 is at AGP2x DIME.

Not that some haven't had problems with Tyan SS7 boards but there are far fewer than with other brands from all indications.

The observation about the "CD" and CE" versions is also correct as most CE motherboards are much better. Also ALI is worse than MVP3 for problems.

PDP-11, Dec-writer & ZD-11 Terminal Unit, RSTS-OS

13th October 1999, 05:45
Anyone know how to tell which MVP3 chip revision is on the motherboard w/o having to open the case (since I'm kinda lazy :P)
I have the SiSoft Sandra Pro, but am not sure which rev the chip is since it just lits a few serial numbers and stuff for the Via MVP3 chipset. Or anyone know of another program that will tell me this info?

13th October 1999, 12:41
I have Scsi, Nic, audio, UDMA/DMA all working fine on my VIA system. My AGP is 1X, but completely stable so I don't mess with it. I run 24/7 for weeks without any problems. When I do have them, A reboot takes care of it. Except for PD5.3 which give BSOD whenever hardware acc. is used.
So I'm back at 5.25

Epox MVP3G-M, WD 8.4gb, Pio 32x CDROM, LS-120, Adaptec 2940U2W, Yamaha CDRW4416, Pio DVD-303, Diamond S90, 3COM 3c509, HP 712c, HP 6200

13th October 1999, 14:26
rylan, check your mobo book

PIII-500mhz @ 620!
Abit BE6 mobo
128mb pc-100
Mill G400(YAHOO!!!) 32mb @ 305 with MGATweak)
PD 5.30
Maxtor 13.2 gb Uata66 hdd
SB Live!
Winblows 98se
THREE BIG FANS, six little ones and a case that sounds like a turbine engine...doh! :D

13th October 1999, 14:31
I'm not positive, but I think that you might see the chipset revision during boot. See if there is a CD or CE buried in that rather long serial number string that is at the bottom of the screen when the memory is being checked during boot.


13th October 1999, 19:33
Disabling Busmastering IS NOT an option! My Epox board ran peachy until 5.30 drivers came around. Only solution here is to go back to 5.25 until Matrox figures out what they fubar'ed.

This kinda falls under the catagory of 'Woops! We broke it! What did we do last?' I would be a lot more inclined to point the finger at VIA or AMD if the previous PwrDsk drivers had not worked so well. As it now stands, I think the SS7 platform is just a scapegoat for poorly written drivers.

13th October 1999, 19:45
It's kind of weird, isn't it, that every time we get more powerful drivers that push memory/cpu limits on a system, SS7 systems have a hard time with them??? Same thing with PD 4.51....Sorry, I would rather that they continue to push the performance envelope, than to neuter the drivers to make them work on less stable systems. You know, if there are previous versions that work, then use those versions.


13th October 1999, 20:54
1. It should say in your motherboard manual what slots are busmastering, the 2 closest to AGP slot is surely dependent on your motherboard. If you have 3 PCI slots with the third one a shared slot of ISA then it won't be busmastering, but that isn't likely to apply to boards with 5 PCI slots.

2. I've had the same SS7 motherboard for almost 18 months now, never had any hardware problems with AGP, busmastering etc. Any paticular board has to potential to be of a bad batch, doesn't matter if it's SS7 of Slot 1 or Slot A. SS7 users have to do more work to install drivers, hence there are more complaints when they don't RTFM. http://forums.murc.ws/ubb/smile.gif

3. No problems with the 5.30 drivers here, in fact a supposed problem with busmastering and tirtanium got fixed. Sometimes it really is a driver bug.

14th October 1999, 03:10
Himlelf, are you really running a G200 on a VIA chipset board with 5.3 drivers? You'd be the first one I know who has succeeded.

(You did not post your system specs...)

---Addendum... well I see that some of you really seem to be able to use 5.3, VIA and G200 at the same time, which seems a bit weird, considering so many are suffering from the BSODs immediately when entering 3d.

There must be a hidden variable somewhere...


Athlon500/MSI6167/64M(!)/10GIBM/6GSamsung/1GSCSI IBM/AHA2940/CL2xDVD/FastMM2+MJPEG/HPPSScanner/HPPSPrinter/G400DH32M/ZIP100SCSI/Mitsumi2xCD-R/Samsung17GL/MX300/AKA Electric Classic

[This message has been edited by Meek (edited 14 October 1999).]

14th October 1999, 06:02
Heheh.. check mobo manual. Don't I feel stupid now. My VIA Chipset is VT82C598, which I'm guessing is rev. C
As for busmastering, I've moved my cards around to differant slots (except PCI1 next to the agp card.. I'll try that soon) and they all apparently support busmastering.
Epox MVP3G-M rev 1.0 incase anyone is wondering. Thought this may be helpful to someone at come point.

Randy Simons
14th October 1999, 11:00
Rev C? You mean CC? AFAIK that was never on a retail mainboard, and surely not on the MVP3G-M. That one is newer than the MVP3E-M which I have and even my old MVP3E-M had the CD revision.
As for that busmastering, afaik all PCI2.1 compatible slots require busmastering. Well, at least my SCSI card does and it doesn't matter where I plug it in.

14th October 1999, 13:19
Yeah, guess that wasn't it. I got the MVP3 serial number from SiSoft Sandra Pro, and it matches wit that at boot, so here it is (the model info part anyway).


What is this, a rev DC?

14th October 1999, 13:22
Im think you'd be better off opening up the case http://forums.murc.ws/ubb/wink.gif

14th October 1999, 15:23

My machine specs are on my homepage, listed in my profile. I have a G400 though. Shouldn't matter when it comes to the drivers working or not since they both use the same archive but it could be G200 specific, no idea. I've had a Riva 128, TNT 1, some misc cards I tested for other people in the thing, and the G400 I have now, now hassles at all, other than Windows being a pain. I've had a K6 200 @ 233, K6-2 266 @ 280, K6-2 350 @ 392, and now a K6-3 400 @ 448 in the same motherboard. Of course it would probably have been cheaper to go with a Celeron and a slot 1 motherboard, but each cpu paid for the next one up to around 50% aprox so no biggy.

21st October 1999, 19:09
I just installed my new G400 32mb DH (not max) and the 5.30 drivers. No more lockups. The problem is G200 specific. AGP is at X2 now also.

Matrox G400 32mb DH, Epox MVP3G-M, WD 8.4gb, Pio 32x CDROM, LS-120, Adaptec 2940U2W, Yamaha CDRW4416, Pio DVD-303, Diamond S90, 3COM 3c509, HP 712c, HP 6200