[LAU] jack_connect and time (batch connecting)

[Joel Roth]

On Tue, Oct 19, 2010 at 01:25:23PM +0200, Julien Claassen wrote:
> Hello everyone!
>   I was wondering about this. In Nama it is allowed to connect
> multiple jack_ports to one ecasound port. So I can connect a pack of
> hydrogen ports to one ecasound port. So I don't have to record
> 30-odd tracks for one drumkit. But the connection takes very long.
> Would it be possible to move all the connects in the background and
> let them finish their jobs there.
>   The number of ports I'm connecting is around 60-66, so 30-33
> pairs. Or is there another way (with external utilities or a Perl
> module, to accelerate this process?

Nama could easily be modified to start 60-or-so parallel
jack_connect processes in the background, however I'm
curious: is there any special reason why jack_connect should
take a long time to return?

Would parallel jack_connect processes be competing for resources
in some way?

It might be informative to start Ecasound in the --server
mode with the -ddd (debugging) flag, and look for any
signs of bottleneck on that end.

Since JACK has an "industrial strength" design, I would
be surprised if the JACK server's implementation turned out
to be a limiting factor.

jack_connect, on the other hand, might suffer from some
sub-optimal behavior.

It would be great if we could track down the reason for
this slow response connecting Ecasound ports to Hydrogen.

Joel

>   Kindest regards
>            Julien


-- 
Joel Roth