On Tue, 2004-01-27 at 17:22, Paolo wrote: > Emiliano Brunetti said: > > On Tue, 2004-01-27 at 17:03, Paolo wrote: > >> Emiliano Brunetti said: > >> [...] > >> > I googled a bit and didn't find much about benchmarking normal desktop > >> > usage. Maybe the case is that there isn't still anything. > >> > >> You don't find anything because you don't know what u want to benchmark > >> ;) > > > > Please read previous posts. I need to benchmark 'normal desktop usage', > > that in such case could be disk I/O, graphic adapter performance and cpu > > load, all in normal conditions - that is somehow simulating a normal > > desktop usage (spreadsheet, writing, surfing, emailing, doing a little > > of graphic stuff). > > It was not my intention to be rude, sorry about that. Never mind. My fault. This is a tough day. Sorry. > May be I just don't understand what you need, > if you want to measure the usage of the machine _while_ you use a few > application then vmstat and top will be your best friends. > If you want to simulate the desktop usage then I'm totaly lost. No, this is not viable. I can't be always monitor all those stuff. I was looking for some kind of tools, even many, to get a overview of system performance as a desktop system, and be able to *repeat* the measure in different conditions. So that i could get a even a very rude view on what is going on by comparing numbers (raw numbers wouldn't be very meaningful...). > My suggestion is to manually use the 'common' application > and monitor the system usage. Well, it can't be a solution if you should be benchmarking several boxes... :( > I really don't know of any benchmark that can be used to measure > the "speedines" of a graphical enviroment. > > I recall it was evend discussed on lkml but I don't think anybody > had a solution to this problem. I see. Thanks a lot, at least i know that my 'art of googling' hasnt' gone nuts. ;) Ciao E.
