Description
1. (30pts)
Under what circumstances does a multithreaded solution using multiple kernel threads provide better performance than a single-threaded solution, on a single-processor system? Describe an example to support your answer.
2. (30pts) How many user-level threads will be constructed with the “pthread_create” API when the program below is run? Explain how you get your answer.
#include <pthread.h>
int i;
void *th_code(void *x)
{
pthread_t tid;
i = i – 1;
if(i>0){
pthread_create(&tid, NULL, th_code, NULL);
pthread_join(tid, NULL);
}
}
int main()
{
pthread_t tid;
i=3;
pthread_create(&tid, NULL, th_code, NULL);
pthread_join(tid, NULL);
}
3. (40pts) Read the following code and answer questions.
#define _GNU_SOURCE
#include <stdio.h>
#include <sched.h>
#include <unistd.h>
#include <stdlib.h>
#include <fcntl.h>
int var;
int do_1(void *);
int do_2(void *);
int main( )
{
void *child_stack_1;
void *child_stack_2;
int pid;
var = 5;
pid=fork();
if (pid > 0) var=1;
child_stack_1 = (void *) malloc(1001);
child_stack_2 = (void *) malloc(1001);
child_stack_1 += 1000;
child_stack_2 += 1000;
printf(“var = %d\n”, var);
if(var = = 1)
clone(do_1, child_stack_1, CLONE_VM, NULL);
else
clone(do_1, child_stack_1, 0, NULL);
clone(do_2, child_stack_2, 0, NULL);
};
int do_1(void *x)
{
var = var*2;
exit(0);
};
int do_2(void *x)
{
var = var*3;
printf(“Thread %d: var=%d \n”, 2, var);
exit(0);
};
Questions: What are the outputs of the above program?
Here we assume the CPU scheduler uses the FCFS scheduling algorithm to schedule the running of threads [Note: fork() creates a process, which is actually the first (and the only if the no more thread is created in the process) thread of the process; hence, the CPU scheduler actually schedules threads].
Explain how you get your answers.
