The Threads Library ``multiplexes''
threads among the pool of available LWPs for the process.
An LWP can pick up and run only one thread at a time.
After a time the LWP will ``put down'' (stop running)
its current thread and pick up another.
Some time later the thread will be picked up again;
not necessarily by the previous LWP.
The algorithm by which a thread is associated with an LWP and later preempted
is covered in
``Multiplexed thread scheduling''.
On multiprocessor systems,
a larger number of LWPs implies a greater chance that
different threads of the process will be executed simultaneously
(that is, true, not logical concurrency).
Managing threads concurrency
Lightweight processes and threads concurrency level
© 2005 The SCO Group, Inc. All rights reserved.
SCO OpenServer Release 6.0.0 -- 02 June 2005