hi Fabio,
sorry to hear about your troubles in the performance.
Honza Svasek had similar errors with ImproSculpt under Linux,
it seems it has to do with the large number of FLTK widgets not running in it's own thread. This will most probably be fixed in csound 5.
But there's still some opcodes that does not work in csound 5,
so it is not possible to run all of ImproSCulpt under cs5 yet.
Parts of it do work though, so you can give it a try if you're curious.
I will release a new ImproSCulpt version when csound 5 is ready.

best
Oeyvind

> From: Fabio Barbon 
> Subject: [CsndTek] RT Csound + GNU/Linux issue
> To: csoundtekno-N4abDuUB7xo@public.gmane.org
> Message-ID: <415126FE.9080302-KfUzieivrh8@public.gmane.org>
> Content-Type: text/plain; charset=us-ascii; format=flowed
> 
> Hi, this is to share an unhappy experience I had with csound and 
> GNU/Linux in a realtime setting.
> 
> During our last live performance (>200 people attending) my laptop 
> freezed twice while running Csound (plus, Csound "segfaulted" once). 
> When this first happened we pretended the sudden silence was a wanted 
> effect, but the following hangs put us really closed to panic.
> 
> I used fltk'ed csound v. "4.24.1" (Istvan Varga patches) for ALSA audio.
> Kernel and ALSA are those from Planet-CCRMA (over RedHat 9).
> Audio hw is M-Audio Quattro, laptop is a "standard" Dell P4-HT 3.2Ghz.
> 
> I use csound as an effect processor (mixer fx send to csound audio in, 
> csound audio out routed back to mixer fx return), so both "ins" and 
> "outs" operators occurs on the same instrument.
> 
> I say that because I've never experienced any problem with Csound using 
> it as an audio "input only" or "output only" machine (and without 
> graphic widgets).
> 
> The computer freezes occured while using beautiful Oeyvind Brandtsegg's 
> ImproSculpt. This huge patch uses fltk widgets heavily (expecially 
> vu-meters are resource demanding, I guess), so I suspect the problem is 
> somewhat related to xwindow event handling (i.e. a XFree86 video driver 
> issue) conflicting with real-time priority ("--sched") for audio, 
> together with some issue of snd-usb-audio ALSA driver.
> 
> I'd like _very_ much to understand what went wrong, and possibly 
> contribute and solve this problem, but so far I really don't know how to 
> face it.
> 
> Any help?
> 
> Many thanks in advance,
> - Fabio
> 
> -- 
> Fabio Barbon
> Automated Reasoning Systems - Formal Methods Group
> IRST - Istituto per la Ricerca Scientifica e Tecnologica
> http://sra.itc.it/people/barbonfab