Hello all,
I am looking for a little bit of explanation on how the X11 session is delivered to the X2GO client; and whether improving the performance of the local X11 server would improve the overall experience?
As a reference point I ran x11perf while connected to a remote X2GO session with the following results:
x11perf -copywinwin10 x11perf - X11 performance program, version 1.2 The X.Org Foundation server version 60900000 on :54.0 from uxbod Thu Apr 29 15:18:53 2010
Sync time adjustment is 0.0263 msecs.
36000 reps @ 0.1401 msec ( 7140.0/sec): Copy 10x10 from window to window 36000 reps @ 0.1460 msec ( 6850.0/sec): Copy 10x10 from window to window 36000 reps @ 0.1457 msec ( 6860.0/sec): Copy 10x10 from window to window 36000 reps @ 0.1424 msec ( 7020.0/sec): Copy 10x10 from window to window 36000 reps @ 0.1450 msec ( 6900.0/sec): Copy 10x10 from window to window 180000 trep @ 0.1438 msec ( 6950.0/sec): Copy 10x10 from window to window
when I run the same test on my physical laptop which has a NVidia Quadro card I get:
x11perf -copywinwin10 x11perf - X11 performance program, version 1.2 The X.Org Foundation server version 10604000 on :0.0 from cyborg Thu Apr 29 15:19:18 2010
Sync time adjustment is 0.0979 msecs.
2400000 reps @ 0.0020 msec (492000.0/sec): Copy 10x10 from window to window 2400000 reps @ 0.0021 msec (486000.0/sec): Copy 10x10 from window to window 2400000 reps @ 0.0020 msec (502000.0/sec): Copy 10x10 from window to window 2400000 reps @ 0.0020 msec (488000.0/sec): Copy 10x10 from window to window 2400000 reps @ 0.0020 msec (496000.0/sec): Copy 10x10 from window to window 12000000 trep @ 0.0020 msec (493000.0/sec): Copy 10x10 from window to window
Reading one of the NX documents about slow performance it mentions:
"Upgrading to a more recent version of the X server or trying a different video driver can surely offer a great boost. For example a hardware accelerated video card is recommended to obtain the best speed."
Would one of the devs be able to cast some light on where I should look to tune the performance?
Thanks, Phil