(I’m using what is probably the most recent Haiku driver for this, but on R5)
I’ve got an odd problem, where the Matrox driver is selecting my secondary output on a G550 PCI-E as BeOS’s primary - this means that with a single monitor, I can see either the BIOS and bootman; or see BeOS.
None of the Dualhead Setup options seem to be of complete use - Switch Head does switch heads but leaves me with an extended desktop across a non existent monitor; Clone Head leaves me with the cloned desktop on my primary output - so no cursor due to the driver using a hard cursor on monitor 1.
Can anyone remember if they had this before and if there is a fix? Other than wiring the second output up and switching between them on the monitor, which I could do but is rather faffy!
From memory: the second head hardware has no hardware cursor capabilities. In the matrox.settings file set hardcursor false and reboot te get a software cursor.
I am not aware of the switch problem, though DVI output is not supported, you need to connect via VGA adapters.
Might be there is an option in the settings file to invert output selection, don’t know by heart.
Since you are on R5 you will be using matrox dualhead setup if all is right, as the screen prefs panel.
Hmm, ok. I have no active memory of that specific part. The driver works well on all of these cards though. If you look at the hardware and if a connector has a internal cable to the board, then I’d say this is the secondary output. Max 1600*1200 resolution I think. And no hardcursor as I said
I also still have a quad G200 which also works well. I used to play some videos on one of the secondary cards while having the desktop on another output (media videonode).
This is not a quad head card but actually four seperate cards on one printed circuit board (4* 8Mb RAM).
I was using the DualHeadSetup for testing - I’ll try disable the hardcursor and see if mirroring works around the problem for me.
Connected using a DVI-VGA adapter - DVI direct not working shouldn’t be a problem if all I want to use the DVI input on the monitor for is viewing the BIOS/bootman on the primary output and then switch back to VGA for the secondary output that BeOS is using as its primary.
I’ve noticed that whatever way it builds the mode list is giving a max res beyond what the monitor can do; but missing the monitors genuine max (showing 1920x1440 when its a 1920x1200 monitor).
Its very hard to find the nVidia PCI-E cards that have support from your nvidia driver; whereas the G550 PCIE is a bit easier to find (in Europe at least) so that’s why I’m using one. Computer is too new to have AGP slots!
Ah yes. So are you really using R5? Did you compile the driver for Beos yourself then? Or did you fetch it from my old site?
(Be-hold: (open)BeOS MATROX MGA driver)
You’ll also find the settings file docs on that site for instance btw. If I remember correctly EDID is not yet implemented in that driver so that’s a reason I expect the resolutions differ. Also the hardware has limited capabities resolution wise, when I worked on this driver a 1920x1200 monitor was nowhere to be found in my neighbourhood at least.
Concerning nVidia I am aware it’s hard to find PCIe supported cards, this is a reason why I am trying to add a few more cards to that driver. Since the hardware changed a bit more since the 79xx cards (testing 8400-9200 atm) it won’t work yet though. I have partial succes with detecting EDID from screens and a small part of modesetting but not useable enough yet.
If you have normal PCI slots still in your system you could try to find a PCI nVidia card, though I/O will be slow of course. Otoh on R5 you will have hardware acceleration for 2D which should make even a PCI card work reasonably well…