On Fri, 2004-07-30 at 23:30, Benno Senoner wrote: > It's probably because you use X11 over ssh. You would think so, wouldn't you? (well I certainly did before I started measuring things). Setup: fvwm and all apps remotely over ssh with blowfish encryption. Remote machine is 1.3GHz Duron c/w 512Mb RAM. Terminal is PII 300(ish) MHz c/w 256Mb RAM. Running apps: Qjackctl, Alsaplayer, Ardour (not recording - just showing a meter), and Meterbridge. With Alsaplayer eye-candy turned on, it's bar graphs move _significantly_ faster than Meterbridge's. Ardour's meter looks 'normal' (i.e. adequately damped), but Meterbridge looks very poor by comparison. Comparatively, Meterbridge is responsible for blowing _loads_ of CPU cycles (as reported by Qjackctl) - Alsaplayer/Ardour graphics aren't loading the remote box to anything like the same extent. I'll try to make some comparative CPU measurements whilst recording and report back, but it looks to me as though Meterbridge (probably via a library) isn't making as efficient use of X as it might. cheers Rob
