Was asked about it a few days ago, and trying to find some proper
material I'm still sort of in the dark.
How does Linux handle multiple processors?
Now I'm not sure about the questions I want to ask - but lets try.
Does the OS fork / create extra jobs to handle the extra processors?
How much of the code is "shared".
How much extra memory is needed? ie. Does each task only require
the non-shared amount of memory needed?
How does the kernel schedule/run tasks on the different cpus?
I have a fair idea how IBM does it on AIX. But Linux seems
to be way different - or am I reading this wrong?
Is there a straight forward block diagram somewhere that illustrates this?
Been to a few sites on the net - but can't find one with all the ins and outs.
I actually have more questions - but cannot think how to phrase them,
and make sense
Helped a client install RHES 2.1 on a dual processor box (first time
fo me that was) - but cannot really make sense of what now runs
and what is needed.
So lets say I do a real basic install of Linux that uses say 64MB of memory
and then slap in an extra processor and do the reinstall...
how much extra memory is used and by what, which extra
tasks should be visible etc. etc.
Know this all really vague, but like I said, I'm not sure about the
questions. So don't be to harse with the answers.
Regards.



