CS411/511. Operating Systems
Spring 1999

Project #0: Introduction to the OSP Simulator

Due at the start of class on Friday, April 9.

Turn in your writeup and the table showing the results of the 5 simulation.run files to cs411@engr.orst.edu. Send it all as a single simple email message (no attachments or special formats).

Follow these steps in learning to use the OSP Simulator.

  1. Log into one of the "state" engr workstations (i.e. md.engr.orst.edu, hi.engr.orst.edu, etc).
    OSP will NOT run on non-HP machines (like flop).

  2. To see the OSP files set up for the demo, use ls:
         > ls ~cs411/osp/demo
    To run the demo, use the following command from your own directory:
         > ~cs411/osp/demo/OSP.demo
    This way of executing OSP requires that you supply the system parameters interactively -- use the responses suggested in the following table.

    intensity of process creation 5
    average CPU time 60
    frequency of events 7
    memory-related events 1
    I/O related events 7
    rsrc-related events 1
    socket type s
    CPU time quantum 20
    prepaging 0
    memory reference pattern l *
    handling deadlocks a
    simulation time 40000
    snapshots 1
    trace switch 0
    interactive n

    * Note that this is the letter l, not the number 1.

    Your responses will be recorded in the file simulation.parameters. Output will be directed to the file simulation.run as long as you answered "n" to interactive mode; otherwise they will be pritned on the screen. The two simulation file names are used over and over, so copy the information to another file whenever you want to keep it.

  3. For subsequent runs, you can use the same parameters without having to type them in again. Copy simulation.parameters to a file called demo.parms. The following command (typed from within your directory) will use the same parameters in a new run:
         < ~cs411/osp/demo/OSP.demo demo.parms
    Save the file simulation.run as demo1.run so you can refer to it again. Use the values to fill in the column demo1 in the table.
    				demo1	demo2	demo3	demo4	demo5
    system throughput
    total number of tracks swept
    average number of tracks 
      swept per r/w request
    average turnaround time of 
      I/O request

  4. To observe how the results change as the parameters are varied, copy file ~cs411/osp/demo/demo.parms into your directory and try a run using these values. Compare the results with those from step (2). Which parameter settings changed? What affect did this have on the statistical results? Save the simulation.run file so you can refer to it again. Record the values it shows in the demo2 column.

  5. Now run the program again in exactly the same way as in step #3 above, 3 more times. Record the results in the demo3, demo4, and demo5 columns. You will probably get different results each time, because the simulation takes into account the fact that the job mix is never exactly the same.

  6. Examine the simulation.run files and try to explain the circumstances that arose causing the variations.

Questions should be sent to the class maillist, class-cs411@engr.orst.edu.