VOGONS


First post, by Nexxen

User metadata
Rank l33t
Rank
l33t

It always made me wonder, the BIOS must be tailored or just having it "translate" (I mean the output data computed by the main chip) signals to other parts is enough to display?
I'm referring to old cards, not modern ones. But if you want to talk about AGP and PCIe, why not.

Feel free to better rephrase my question if necessary, I know it is very badly put.

Thanks

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

"One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios

Reply 1 of 4, by mkarcher

User metadata
Rank l33t
Rank
l33t

Many video chipset have configuration parameters for the bus interface, for example to configure the amount of waitstates. These parameters work differently depending on what bus the chip is connected to. It is quite usual to that the initialization code in the video BIOS needs to perform different steps depending on the bus type. On the other hand, having less variants of a video BIOS makes version management easier for the chip manufacturer, so it is quite common to see "universal BIOSes" that detect the bus type and initalize the chip accordingly.

Another challenge is that the bus interface of older video chips often isn't "glueless", i.e. you need some interface chips between the actual bus and the video chip. Sometimes, the correct initialization depends on how the glue logic is implemented on the video card, and customizations like this which are performed by the card vendor might tie a specific BIOS variant to a certain card, so it fails to work on other cards with the same chipset.

As video cards got higher integrated over time, there are less variations between cards, so for AGP/PCIe, I expect most BIOSes to work universally. Nevertheless, a BIOS for a "overclocked by vendor" card will likely try to apply the same overclocking if placed on a different card, which might fail to work.

Reply 2 of 4, by Nexxen

User metadata
Rank l33t
Rank
l33t
mkarcher wrote on 2024-07-08, 14:39:

Thanks for the reply. Learnt something.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

"One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios

Reply 3 of 4, by akimmet

User metadata
Rank Member
Rank
Member

There are many early SVGA chips that would work with multiple busses depending on configuration. It looks like many determined bus type by strapping certain pins to ground or Vcc.
I never tried swapping a VL-Bus BIOS with a PCI BIOS for the same SVGA chip to see what would happen. I have always assumed they are different.

Reply 4 of 4, by mkarcher

User metadata
Rank l33t
Rank
l33t
akimmet wrote on 2024-07-08, 15:48:

I never tried swapping a VL-Bus BIOS with a PCI BIOS for the same SVGA chip to see what would happen. I have always assumed they are different.

I happen to know that most Trio64V+ BIOS chips do read the straps, and at least one regster is programmed differently depending on the bus type. The BIOS contains two vendor-customizable values to initialize a specific register. The first value is used for VL cards and the second value is used for PCI cards.