I read this in the jackit-devel list archives:
http://sourceforge.net/mailarchive/forum.php?thread_id=9639236&forum_id=3040
I have an Athlon 64-X2 machine, which I'd intended to use at the
main machine in my studio. I also have a Celeron machine, which
is still in use, having started out as my proof-of-concept for
a Linux studio box. One person commented that with their 64-X2
machine, that a larger period size, 256 versus 128, seems better.
This is generally true for avoiding xruns.
I'm wondering, is an even larger period size 1K or 2K going
to make a 64-X2 machine usable. I primarily use ardour,
rosegarden4, and audacity for my audio work. What factors impact
the timing problem that I might exploit to minimize it? I'm
not too adverse to trying newer versions of a kernel and jack.
Right now I'm running kernel 2.6.12 jackit 0.100.1. I remember
something about there being a pervasive udev change starting in
the 2.6.15 kernel. I'd like to stay under that rev., if that's
true.
Related to this is that I'd like to retire my 200MHz Pentium
I machine in our office - 10 years old this month. Would you
recommend avoiding the 64-X2 machine for now with my audio work
(OK, so it's play)? If so, then I'll keep using the Celeron for
music, and move the 64-X2 into the office. We'll be able to
compose emails faster than anyone on the block; sigh. It sounds
like I really just need to be patient and let the kernel & jack
development teams sort this all out. But, if I can make some
adjustments and get useful results, then I'd move that direction.
Thanks....
P.S.: If you think I should subscribe to jackit-devel and take
my question there, then I'll do that.
-- KevinReceived on Sun Feb 26 20:16:36 2006
This archive was generated by hypermail 2.1.8 : Sun Feb 26 2006 - 20:16:36 EET