Re: Successful ruby setup and a problem

[Helge H. <hel...@ai...>]

Boszormenyi Zoltan wrote:
> Helge Hafting írta:
[...]
> 
> dmesg reports now two DRI devices, the old DRM driver did not report
> anything. Minor 0 is the PCI Radeon, minor 1 is AGP Radeon.
> My first attempts setting up DRI on both cards failed.
> I used the AGP Radeon for :0.0 and the PCI one for :1.0,
> this did not work. Somehow the detection code cannot work out
> this config, it tries to set the "private label" of the DRI device
> to the PCI bus id of the device and it fails somehow.
> Changing the display order makes DRI work on both!
> 
I have a matrox G550 AGP card, and I'm trying to figure out what PCI
card to get, that will support DRI.  Looks like the Sapphire radeon 7000
is a good choice.  :-)

>> Also, it looks like you have a single cpu.  Are both games nice
>> and smooth with such a setup, or does the first game get
> 
>  > noticeably worse when the second one starts?
> 
> They both show occasional hiccups (some frames lost) but they
> are very fast. Yes, a single CPU but such a beast! :-)
> Running at 2GHz with an 1MB L2 cache. Did I say Athlon64 3200+?
> I think that the scheduler tweaks that made their way into
> 2.6.4-rc2-mm1 would help here.
> 
Good to know.  I'm planning to upgrade from the current dual
333MHz celeron, and dual opteron is kind of expensive right now.
I want something 64-bit capable though.  

>> I've noticed that tuxracer doesn't like a compile going on at
>> the same time, but that's on a 333MHz processor. A dual cpu
>> setup fixed that for me.
> 
>>> My main problem is that localhost:1.0 is destroyed by logging
>>> out on localhost:0.0.
>>
>> Destroyed how - logging out kills the other xserver too?
> 
> The monitor goes black, looks powersave mode but it's not.
> Pressing Ctrl-Alt-BkSpace brings back the gdm login.

I see something similiar, my current setup uses the matrox driver
for the first screen and the framebuffer driver on the second
framebuffer on the same card. 

Switching consoles on the first screen (or starting/stopping X)
always switch mode on the other screen too.  I always thought the
problem was because it is a single card and features for the
two displays aren't well separated in the driver because "everybody"
uses it for dual-head single-keyboard setups where it doesn't matter.
But now I'm not so sure.  I fixed the login/logout problem by
running a matrox-specific program that resets the mapping between
framebuffers and vga connectors each time X runs.  That fixes
most of the problems for me.

I have also noticed that something goes wrong with console
blanking or some such.  Leaving the computer alone overnight
occationally gives me a black screen on the framebuffer screen,
even though X is running there.  X is unaware of it and doesn't
repaint automatically, but moving a window around fixes that.

Helge Hafting