> I am running a 2.4 kernel, yes. I seem to recall that in 2.6 the
> clock does run 10 times faster than in 2.4; perhaps muse contains
> of hardcoded assumption about the speed of a timer somewhere. I'll
> if I can have a look along these lines tonight. Excellent thought.
Sorry, I also missed that you had actually written that you were using
I had a peek at the lines (think it was in dummyaudio.cpp, perhaps
alsamidi.cpp was involved), and IIRC the calculation did not take the
actual timing resolution into consideration. But it was only a short
peek and I didn't do any experimentation. I think Werner has made a new
dummy audio driver in the 0.8-branch, actually, I haven't tested that
one either (or did I? I don't remember...).