I have a doubt in Multi Tasking.
Suppose if the process makes an I/O request at 5 ms in a 10 ms time slice, wouldn’t it be more efficient if the scheduler detects this and stops the process immediately and ‘gives’ the cpu to the next process and restarts the time interrupt (time slice)? In this way the remaining 5 ms can be utilized. The unproductive time of cpu included in this remaining 5 ms will leave us approx 4.something ms.
YES. This is what the scheduler does !! The process that has made an I/O request cannot continue anyway ( since it needs the data to come for processing ) and the next process to be scheduled would surely like to have the full time slice.
Please login first to submit.