On Sat, Sep 29, 2001 at 11:55:31PM +1000, Scott Howard wrote: > applications where it suits (and it doesnt always suit). The overhead of > a context switch is always going to be greater than that of switching > between threads (which all occur within the same context).
I was under the impression that a different thread *is* a different context... it has its own stack, its own registers, and probably can have its own memory address space too. They're basically like processes, except they tend to share an awful lot of their memory with other threads of the same "process". In this case, I fail to see why it would be significantly faster to switch between threads than it is to switch between processes. I should probably write a program to benchmark this to find out for certain, I suppose. I could well be incorrect about all this... if so, I would greatly appreciate it if someone could please enlighten me. Regards, Andrew. -- SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/ More Info: http://lists.slug.org.au/listinfo/slug
