GATE Computer Science Engineering(CSE) 2027 Test: Introduction, Process,

The only state transition that is initiated by user process itself is block.

In pre-emptive scheduling suspanding temporarily a running process is done before the CPU time slice expires.

The Round Robin CPU scheduling technique will become insignificance if the time slice is very small or very large. When it is large it acts as FIFO which does not have a fixed determination over the turn around time. If processes with small burst arrived earlier turn around time will be less else it will be more.

Considering that a process spends a fraction p of its time in I/O wait state. As n processes are there is memory at once, then Probability that all n processes are waiting for I/O = p x p x p x ... p (n times) = pⁿ
Note: Moreover, the CPU utilization is given by the formula CPU utilization = 1 - pⁿ.

The arrival pattern is a Poission distribution.

Context switching takes place when a process is preempted (forcefully) and another process goes into the running state. In FIFO and SJF (non- preemptive) techniques, the processes finishes their execution then only their context is switched to other processes.

In preemptive scheduling the process with higher priority is executed first and then next higher process with least priority is executed after the higher priority processes are executed, if occured simultaneously. This is the ideal technique to be used in real time O.S. in which critical or higher priority processes are to be entertained first.

SJF will gives minimum average waiting time. SJF ensures that all smaller processes will finish first. So, less processes will be waiting since more smaller processes will end up in less number of time. Use processes waiting means less waiting time overall which will decrease the average waiting time.

A fork () call creates the child processes for the following loop creates 2ⁿ - 1 child processes, excluding the parent process. And including parent it is 2ⁿ.

Kernel level threading may be preferrable to user level threading because storing information about user level thread in a PCB would create a security risk i.e. with each access to the non-critical services. We are going to the domain where both critical and non-critical services are residing. Any harm in this domain may creates problems to the critical services. Hence Kernel level threading is preferable.

Only critical services must resides in the kernel. All services mentioned except reading status of processors and reading time of the day are critical.

(i) is false. Context switching between user-level threads is generally faster because it is done entirely in user space and does not require a switch to kernel mode or invocation of the kernel scheduler.

(ii) is true. If a user-level thread makes a blocking system call, the kernel sees the whole process as a single schedulable entity and the blocking call can cause the entire process (and thus all its user-level threads) to block.

(iii) is true. Kernel-supported threads are visible to the kernel and can be scheduled independently by the kernel scheduler.

(iv) is true. The kernel is unaware of the presence of multiple user-level threads inside a process, so user-level threads are transparent to the Kernel.

Therefore, statements (ii), (iii) and (iv) are correct and option (A) is the right choice.

• User level threads are not recognised by operating system, hence do not need any hardware support. Since, user level threads are not recognised by operating system, no separate PCB (Process Control Block) information is to stored. Hence, context switching is faster.  
• Since, Kernel level threads are like separate processes. So, blocking one thread will not block other threads.

Only FCFS and non-preempting algorithms favour CPU bound processes.

It all the processes are I/O bound, they will go or reside into BLOCKED state and hence READY QUEUE becomes empty.