Billy Holmes wrote:
This makes it easier and faster for the kernel to context switch
between processes (and threads within a process).
A prime example of this is a test from the LKML, where Ulrich Drepper
wrote he was able to start and stop 100,000 threads in about 2
seconds. In further messages, Ingo Molnar confirms this and also
points out that before these changes, it would have taken upwards to
15 minutes for the same thing.
http://kerneltrap.org/node/422
Absolutely true. NPTL is drastically more scalable than linuxthreads.
For instance, spawn 1000 worker threads in a JVM and the scheduling
overhead alone will cause the system to become unresponsive across the
board. Use NPTL and you can get 10000 threads all running no problem.
-ryan
--
[email protected] mailing list
