I've read a bit in the technical PDF documentaion for Voodo1 and Voodo² cards wich i downloaaded from falconfly.de. The way to program the cards looks very strange. It has only a linear adress-space configured by PCI-BIOS. There is no IRQ or IO, all io, frame and texture buffer are decoded in this memory area. If there are no bugs like the famous S3-bug and the PCI BIOS is configured right there are no adress-problems like the famous soundcard desaster (but ISA cards are another story). By the way, what is the IRQ on graphic cards for? The Voodoo1+2 have none and my ATI has a jumper for activate and deactivate the irq. I checked the speed with 3DMARK 99 with IRQ on and off but there was no difference!

Regards,
Andreas

That's one of the things I love about the Quantum3D drivers, patience.  The way they put the X-24 under "Secondary Display Adapters".  Much better! ;D

And, yes, nudgegoonies, I am always amazed at the tidy way the Voodoo Graphics and Voodoo2 were implemented in hardware terms.  Which makes it all the more ridiculous that Microsoft dropped formal support for add-in 3D accelerators with Windows NT/2000/XP.  >:(

I don't understand why the graphics card's IRQ is optional for some cards and drivers (it isn't always the case).  It may be hardware specific.  But giving something as important as the PC's picture output an IRQ (hardware interrupt) makes more sense than some of the peripheral rubbish that seems to demand one! ;)

..the appearance of the voodoo cards in the multimedia&gamecontroller section looks strange at first look. But i think it is only a result of the drivers because they only add properity drivers like glide or additional functions for directx and don't work as a full graphic-card driver. That brings me to the question why 3dfx didn't made a BIOS with VESA for the card. How much is a Voodoo 2 missing from a real grahic card? A VESA BIOS, some 2D functions like STRETCH and a FONT cache for GDI, then the right drivers and thats it! Or am i wrong? Or were the cards first usage for CAD stations where they always have the Program on one Monitor and the 3D output on another? ???

By the way "right drivers". The PDF's about the Voodoo1 and Voodoo2 are very complete (are PDF's for Voodoo3 and VSA-100 also available). I think the way to code a driver with the DDK and full support for DIRECTX is such a hard task that no freak at home is able to do the job. Also most unofficial driversets, either the Athlon- or SLI fix for Voodoo2 or RAGELTMAN's ATI drivers, they all put different versions of driversets together and manipulate the INF file. Some files are hacked but is there a real driverset out there that was coded from scratch?

Regards,
Andreas

The Voodoo2 is amazingly close to being a "full" graphics card.  The beta 3dfx drivers for Window NT prove that.  With that driver the desktop is displayed by the Voodoo2 and the Voodoo2 behaves just like any other graphics card (although there are some display anomolies...).

I'm not sure why 3dfx didn't go for a single 2D/3D card implementation for Voodoo Graphics and Voodoo2 except that the chipsets were really for 3D only arcade gaming machines.  The Rush and Banshee didn't quite hit the mark, did they?  And the add-in idea worked really well until Microsoft changed the rules.  >:(

And I've yet to see any driver created by "outsiders" that wasn't just a slight make-over of an existing driver.  I too doubt that any group of private individuals have the knowledge and resources necessary to create such complex things from scratch.  It's simply too big and tricky a job.  Even the big boys (nVidia, Matrox, ATI) can't get them right! ;)