> It's all in the accounting. Use usleep(0) if you want it to "look good".
Two things:
1. First, I think there's a misunderstanding on what my
original issue was: I am not interested in any way by
CPU cycle accounting, and wether the yielding loop should
log any of it. All I want is: when I run a bunch of
yielders and a actual working process, I want the
working process to not be slown down (wall clock) in
anyway. That's all. What top shows is of little interest
(to me). What matters is how many real world seconds it takes
for the actually working process to complete its task.
And that should not be affected by the presence of running
yielders. And, David, no one is arguing the fact that a yielder
running all by itself should log 100% of the CPU.
2. I have a question about usleep(0). You seem to make the point
that usleep(0) is equivalent to sched_yield(). I can see how
intuitively this should be the case, but I am not sure if it
will always be true. It's certainly documented anywhere.
- Mgix
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Jun 23 2002 - 22:00:16 EST