From: Shane Dempsey (sdempsey at domain tssg.wit.ie)
Date: Wed 29 Sep 1999 - 11:18:27 IST
Sorry to add more to this subject that we shouldn't be bringing up again
but doesn't gnome use the orbit orb for interprocess communications?
That would mean that communications would be carried out over sockets.
I think that orbit has some optimisations for local communications however
and I am not sure how they would fit in with the d11 model that
Caolan mentioned. There will certainly be some overhead but it may
have less to do with the sockets and more to do with using an ORB to
handle window manager communication activities.
I'm sure that someone else on the list has a better insight into this
as I haven't read much on gnome or orbit. That was just my initial
impression on its architecture.
...shane
-- ################################### Shane Dempsey - sdempsey at domain tssg.wit.ie TSSG Researcher http://www-tssg.wit.ie ################################### ----- Original Message ----- From: Caolan McNamara <Caolan.McNamara at domain ul.ie> To: <ilug at domain linux.ie> Cc: <kengu at domain credo.ie> Sent: 29 September 1999 10:25 Subject: RE: [ILUG] Raph Levien Interview > And just to drive ilugers mad by bring this up again... > l) In relation to X speed ups have you any thoughts on an improvement to X's > speed by creating a faster ipc mechanism for the case where the app and the > client are on the same machine. Something along the lines of > http://reality.sgi.com/opengl/d11/d11.html. Perhaps using the door ipc mechanism > to replace sockets for the local case. While Shm can speed up bitblitting for > the localcase, all the X calls are routed through sockets dya reckon something > like this would be beneficial ?
This archive was generated by hypermail 2.1.6 : Thu 06 Feb 2003 - 13:04:37 GMT