Izumi Tsutsui <tsutsui@ceres.dti.ne.jp> writes:

> It looks the unrolling version is a bit slower.
> (Maybe due to calling intr suspend/resume functions)
> I didn't check difference in interrupt response, though.

If you measured "wall clock" time, then any actual interrupts that
are serviced would add to the total time as well.  OTOH, those
interrupts would be queued up and serviced immedlatiely after
G2_UNLOCK in the non-unrolled case, so I suppose it should appear in
those measurements as well...


  // Marcus