[LAD] Linux Audio 2012: Is Linux Audio moving forward?

[Paul Davis]

On Thu, Oct 11, 2012 at 1:10 PM, Dominique Michel <
dominique.michel at vtxnet.ch> wrote:

>
> In recent years, the biggest improvement was jack2 on multiprocessor
> machines. It is just much more easier to get the job done because the
> processor use is lower. This give us also a better stability at high
> system load.
>

I often get things wrong, especially on #jack on IRC. But it irks me to see
people continue to make this claim about jack2. Let me reiterate: the
normal version of jack2 does not and cannot use multiple processors if the
data flow is serialized. Mutiple processors are only used where there is
parallel flow between clients. Although this is not unheard of (e.g. two
synths flowing into a DAW), its unlikely to be the common case for most
users.

Many people find jack2 to be "more stable" because it does not zombify
clients in its default mode operation. It allows clients to be "late" by a
substantial amount, which can result in clicks and pops in their "apparent"
audio stream, but doesn't cause the server to kick them out in the way that
jack1 does. One way to look at this is that jack2 is more tolerant of
programming errors and system latency than jack1.

--p
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxaudio.org/pipermail/linux-audio-dev/attachments/20121011/906caf8e/attachment.html>