At 08:25 +0000 09/27/2003, ML wrote:

>>  Do you mean an IMS Ultimate Rez card?  There was no Twin Turbo 3D,
>>  unless they marketed a different card under that name outside the US.

>No idea, I bought it used in germany. OF sees it as IMS,tt3d the chip is
>  labeled 'Twin Turbo 128 3d', the ROM says 'Apple iX 3d 1.01'

That ROM label is found on the Apple OEM Ultimate Rez card.

>>  The Twin Turbo has no 3D capability and works well in the Pulsar.
>>  The chip on the Ult. Rez. card is labeled TT 128 - 3D though.
>ok, then it's an Ultimate Rez.
>
>>  The Ult. Rez. card has problems in the lower (PPB) slots under Mac
>>  OS.  However, I've found this to be a driver issue.
>Yep, I can confirm this. The MacOS driver hangs occasionally if the card is
>  behind the PCI bridge. ( and it hangs reproducable when you try to change
>  resolution )

If you disable the extension (driver) which provides acceleration the 
resolution change freeze goes away.   Of course, the card is 
considerably slower without that driver loaded.

>>  Also, I'm not sure how it affects BSD, but the Apple ROM'ed Twin
>>  Turbo works better under OSX than the IMS ROM'ed Twin Turbo.
>Well, this has an Apple ROM. Gives an OF text console. I'll try the generic
>  kernel again, all newer kernels compiled from CVS were quite unusable...

Only the Twin Turbo card is more reliable.  The Ult. Rez. seems to be 
equally bug ridden whether you have the Apple version or the IMS 
version.

>>  In theory, there should not be any difference in compatibility
>>  between Bandit and PPB slots.  However, there is definitely a bug in
>>  the Apple firmware (ROM) with respect to PCI-PCI Bridges.

>Yep. The Millennium II drivers didn't work at all behind the bridge, a
>  Millennium I did.

The bug seems to be related to daisy chaining PPBs and anything that 
looks like it is a daisy chained PPB.  I do not know enough about OF 
to write intelligently on the subject.   But the problem arises when 
a card bearing a PPB is installed in the lower slots of the Pulsar 
behind the PPB which is on the Pulsar board.   The machine freezes 
while (I think) initializing the PCI cards.  At least, with some 
2940UWs installed which have an LED which is initially lit and then 
goes out, the machine will freeze with the 2940UWs above the problem 
card unlit and the ones below it lit.  I don't know at what phase 
that LED gets turned off.

However, to complicate matters, the problem is not completely limited 
to boards with a PPB.  The VST UltraTek/66 (Promise Ultra/66) also 
exhibits this problem.   In the Device Tree the UT66 card has two 
entries, perhaps related to the two IDE channels though the card 
appears as a single SCSI bus to the OS.

And, of course, there's the Ult. Rez. problem.   Several other video 
cards are problematical, but these are usually dual function cards, 
which are either dual headed, or have both monitor support and video 
capture or video output functions.

Another detail is that if the problem card is the *only* PCI card in 
the lower, PPB controlled slots, it will usually work without 
problems.   Additionally, if there are other cards present, but they 
do not have firmware on the PCI card (ethernet, USB, firewire) the 
problem card will usually function without problems.

Jeff Walther