Oh, maybe I haven't made use of the round-trip delays sufficiently. I understand that it is obvious to use round-trip when many requests are sent in a loop, so we can get the cookies together, then ask for all the replies. But in many situations that a single request is sent, we cannot suppress the round-trips. Is that? Thanks --Jianjun <div>On 11/7/06, Vincent Torri <<a href="mailto:vtorri@univ-evry.fr" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)"> vtorri@univ-evry.fr</a>> wrote:<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> On Tue, 7 Nov 2006, Yang JianJun wrote: > > I have used gtkperf to test the gdk-xcb performance on PC, but the benchmark > data shows gdk-xcb is a bit slow than gdk-x11 :( I've experienced the same thing when I ported evas. It might come from a difference in the cache used by Xlib and XCB. I can't remember what you should do to increase the XCB cache (it's an env variable to set), but it's in the ML somewhere :) Also, have you really taken into account the round trips ? Vincent</blockquote><div>  </div> </div>