Problem $1$: $(8$ pts$)$ User$-$Kernel Threads

General: $(2$ pts$)$ You are writing a multi$-$threaded application that serves many distinct groups of client programs concurrently. Let$$s assume that there are $8$ physical cores, with $4-$way simultaneous multi$-$threading $($SMT$).$ The goal is to take advantage of the underlying parallel processing environment. Let$$s assume that our OS is such that it restricts the total number of software threads to $120.$ Provide a kernel to user threads configuration for the: M:$1,1$:$1$:$,$ M:N model.

Discuss how you will distribute software threads $($kernel and user threads, which model you will opt for, what numbers for threads you will select$)$ in order to increase the response time of your system while you also wish to accommodate special requirements such as:

Case $1$: $(1$ pts$)$ sensitive group applications with special priorities

Case $2$: $(2$ pts$)$ client services including ipc with pipes and shared memory,

Case $3$: $(3$ pts$)$ assume that the kernel runs all tasks on logical cores using a round$-$robin equal quantum scheduler of $4$ msecs for every kernel thread that is scheduled on a logical core. We wish to ensure that if application A $($initiator$)$ executes for X msecs then application B $($follower$)$ will execute for $2$X msecs right after. Application requests make average use of IPC. In that case, what distribution of hardware$-$software threads should I choose? Assume we have $40$ initiator applications. Is it possible to determine the minimum and$/$or maximum time quantum in msecs for X that permits the prescribed execution? If it is$,$ what is that value? If it is not, why not? What is the minimum amount of time to service all $40$ initiator and follower applications?