On fre, 2004-07-23 at 16:56 -0400, Bill Nottingham wrote: > Kenneth Porter (shiva@xxxxxxxxxxxxxxx) said: > > --On Friday, July 23, 2004 8:21 AM +0200 Hans de Goede > > <j.w.r.degoede@xxxxxx> wrote: > > > > >The basic idea is todo things the XP way and loads lotts of (cr)app(s) > > >while the user is already seeing a GUI. > > > > Paraphrasing Knuth, premature optimization is a bad investment. Before we > > start down this road, let's see some measurements showing exactly how much > > time each step in the boot process takes. We can then decide where to > > invest effort to improve with the limited development resources available. > > > > Has anyone instrumented the boot process? For example, syslogging each > > initscript to get a timestamp would tell us how much time was spent in each > > one. > > It's been oprofiled a couple of times... the majority of the time seems > spent in disk I/O. :/ That is what RML says as well, his preload work is an attempt at amending this I think, but I really think it would be wise to look at what he has done in this area. - David
