Re: [linux-audio-dev] linux real-time audio performance

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: Re: [linux-audio-dev] linux real-time audio performance
From: est_AT_hyperreal.org
Date: pe elo    20 1999 - 01:40:38 EDT


Paul Winkler discourseth:
>
> Just making sure I've got this right... You're talking about with the
> POSIX realtime stuff, right?

Yes indeed..sorry I was vague there. :)

> Have you compared this to equivalent code that doesn't use the POSIX stuff?

Not extensively. During early oolaboola development it became clear
that going real-time eliminated the totally reliable dropout each time
top(1) updated and during much X activity (this with a 1/50th second
audio buffer). That was enough data for me. :) Also, It's pretty easy
to see the limits of non-realtime processes a priori. I believe the
default linux time-slice is 210ms. A compute-bound process can
monopolize the cpu for that long unless a real-time process grabs it.

> > > Care to write a paper about it?
> >
> > Well, these answers should be integrated into Paul Winkler's FAQ..and
> > people should keep asking questions. :)
>
> Will do. I'm a bit slowed down with it at the moment. Maybe next week
> some new stuff will be up. the main thing I am working on is adding
> user-level posting/editing capability. There are some cgi packages out
> there that do this; I'm checking them out to see what would work for me.

Neato. :) I really like your FAQ and I think it can be a great
focusing device for multimedia support concerns.

Eric


New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : pe maalis 10 2000 - 07:25:52 EST