[plug] NVidia drivers and RENDER acceleration
Craig Ringer
craig at postnewspapers.com.au
Fri Oct 17 00:02:04 WST 2003
> Explains why a gnome-terminal session takes so long to start up.
Perhaps. I suspect that's more to do with the amount of extraneous crap
GNOME-terminal is linked into, though. Try timing a C compile job under
GNOME-terminal, then under xterm, though - you'll be horrified at the
difference the display of the make output makes. I'm yet to try with and
without RenderAccel but I suspect gnome-terminal will be a /lot/ closer
to xterm when accelerated RENDER is enabled.
Now, if only the free XFree86 drivers (ANY of them) supported
accelerated RENDER. I hear the sis driver might soon, and mga kind of
does, but it's not really a good situation right now. I fear the chances
of the s3 driver getting accelerated RENDER are about nil, and I could
/really/ use that. Hmm... maybe finding someone willing to take a bit of
money to implement it might be possible.
Anyway, this render acceleration finially makes mozilla behave like less
of a slug, which by it's self is a really pleasant change.
> On a slightly different tangent I was running the xfree nv drivers for a
> while and it used to take a good couple of seconds to switch from X to a
> virtual console. It really annoyed me at first because it took so
> appallingly long but I didn't have the time to stuff around with trying
> to get the nvida drivers working. When I finally did get around to it I
> was amazed at how quick the change happened and how I had become
> accustomed to the change.
Interesting. It still takes at least a full second for my system to
switch from X to a VC, and at least 1/2 a second to switch back. OTOH,
at least when switching back that appears to be almost entirely my
monitor re-syncing to the new mode. That'd be one of the nicer things
about LCDs - no long re-sync blanking.
Craig Ringer
_______________________________________________
plug mailing list
plug at plug.linux.org.au
http://mail.plug.linux.org.au/cgi-bin/mailman/listinfo/plug
More information about the plug
mailing list