a. Write (or calculate) an event-driven simulation to help you decide which storage placement strategy should be used at this installation. Your pro- gram would use the job stream and memory partitioning as indicated previously. Run the program until all jobs have been executed with the memory as is (in order, by address). This will give you the first-fit type performance results.

b. Sort the memory partitions by size and run the program a second time; this will give you the best-fit performance results. For both parts, (a) and (b), you are investigating the performance of the system using a typical job stream by measuring:

1. throughput, how many jobs are processed per given time unit;

2. storage utilization, the percentage of partitions never used, the percentage of partitions heavily used, and so on;

3. waiting queue length; 4. waiting time in queue; 5. internal fragmentation.

Given that jobs are served on a first-come, first-served basis:

c. Explain how the system handles conflicts when jobs are put into a waiting queue, and there are jobs still entering the system. Which job goes first?

d. Explain how the system handles the “job clocks,” which keep track of the amount of time each job has run, and the “wait clocks,” which keep track of how long each job in the waiting queue has to wait.

e. Since this is an event-driven system, explain how you define “event,” and what happens in your system when the event occurs.

f. Look at the results from the best-fit run and compare them with the results from the first-fit run. Explain what the results indicate about the performance of the system for this job mix and memory organization. Is one method of partitioning better than the other? Why or why not? Could you recommend one method over the other given your sample run? Would this hold in all cases? Write some conclusions and recommendations.

At one large batch-processing computer installation, the management wants to decide what storage placement strategy will yield the best possible performance. The instal- lation runs a large real-storage computer (as opposed to “virtual” storage, which is covered in Chapter 3) under fixed partition multi programming. Each user program runs in a single group of contiguous storage locations. Users state their storage requirements, and time units for CPU usage on their job control card; it used to, and still does, work this way, although cards may not be used. The operating system allocates to each user the appropriate partition and starts up the user’s job. The job remains in memory until completion. A total of 50,000 memory locations are available, divided into blocks as indicated in the previous table.