Hi When a switch from thread 1 to thread 2 is done, actually there do occure two context switches. One from thread 1 to the scheduler and a second from the scheduler to thread 2.
I wonder why the scheduler's algorithms aren't computed in the contexts of the user threads so that one could switch from thread 1 to thread 2 directly. This would halve the number of context switches. I managed to write a minimal thread library using the pth_uctx functions following the above described strategy and it seems to work (on a Debian GNU/Linux sarge, gcc-3.3) The only case where one needs an extra context is when a thread is quitting as one cannot destroy a context from within this context (can you?). For the case you want to have a look at this proof of concept: http://copton.net/pth-fast.tar.gz Is there a reason for the current implementation I don't see? cu -- Alexander Bernauer ______________________________________________________________________ GNU Portable Threads (Pth) http://www.gnu.org/software/pth/ Development Site http://www.ossp.org/pkg/lib/pth/ Distribution Files ftp://ftp.gnu.org/gnu/pth/ Distribution Snapshots ftp://ftp.ossp.org/pkg/lib/pth/ User Support Mailing List pth-users@gnu.org Automated List Manager (Majordomo) [EMAIL PROTECTED]
