1 Friday, June 16, 2006 "In order to maintain secrecy, this posting will self-destruct in five seconds. Memorize it, then eat your computer." - Anonymous

2 §g++ -R/opt/sfw/lib somefile.c to add /opt/sfw/lib to the runtime shared library lookup path Edit your.bash_profile and add /opt/sfw/lib to LD_LIBRARY_PATH

3 Scheduling in Unix - other versions also possible §Designed to provide good response to interactive processes §Uses multiple queues §Each queue is associated with a range of non-overlapping priority values

4 Scheduling in Unix - other versions also possible §Processes executing in user mode have positive values §Processes executing in kernel mode (doing system calls) have negative values §Negative values have higher priority and large positive values have lowest

5 Scheduling in Unix §Only processes that are in memory and ready to run are located on queues §Scheduler searches the queues starting at highest priority §first process is chosen on that queue and started. It runs for one time quantum (say 100ms) or until it blocks. §If the process uses up its quantum it is blocked §Processes within same priority range share CPU in RR

6 Scheduling in Unix §Every second each process’s priority is recalculated (usually based on CPU usage) and it is attached to appropriate queue. §CPU_usage decay

7 Scheduling in Unix §Process might have to block before system call is complete. While waiting it is put in a queue with a negative number. (determined by the event it is waiting for) Reason: l Allow process to run immediately after each request is completed, so that it make the next one quickly l If it is waiting for terminal input it is an interactive process. §CPU bound get service when all I/O bound and interactive processes are blocked

8 Unix scheduler is based on multi-level queue structure

9  top provides an ongoing look at processor activity in real time  nice default value is zero in UNIX l Allowed range -20 to 20

10 Windows §Priority based preemptive scheduling §Selected thread runs until: l Preempted by a higher priority thread l Time quantum ends l Calls a blocking system call l Terminates

11 Win32 API §SetPriorityClass sets the priority of all threads in the caller’s process. l Real time, high, above normal, normal, below normal and base §SetThreadPriority sets the priority of a thread compared to other threads in its process l Time critical, highest, above normal, normal, below normal, lowest, idle

12 Win32 API §How many combinations? §System has 32 priorities

13 RR for multiple threads at same priority

14 §Selection of thread irrespective of the process it belongs to. §Priorities are called real time, but there are not. §Priorities are reserved for the system itself

15 §Ordinary users are not allowed those priorities. Why? §Users run at priorities 1-15

16 Windows §Priority lowering depends on a thread’s time quantum §Priority boosting: l When a thread is released from a wait operation Thread waiting for keyboard Thread waiting for disk operation l Good response time for interactive threads

17 Windows l Currently active window gets a boost to increase its response time l Also keeps track of when a ready thread ran last l Priority boosting does not go above priority 15.

18 Example: Multiprogramming §5 jobs with 80% I/O wait §If one of theses jobs enter the system and is the only process there then it uses 12 seconds of CPU time for each minute. §CPU busy: 20% of time. §If that job needs 4 minutes of CPU time it will require at least 20 minutes in all to get the job done

19 Example: Multiprogramming §Too simplistic model: l If average job computes for only 20% of time, then with five such processes CPU should be busy all the time. l BUT…

20 Example: Multiprogramming §But we are assuming all five will not be waiting for I/O all the time.  CPU utilization: 1-p n n = number of processes p= probability that all n is waiting for I/O at the same time Approximation only: There might be dependencies between processes

21 CPU utilization as a function of number of processes in memory

22 Producer-Consumer Problem §Paradigm for cooperating processes, producer process produces information that is consumed by a consumer process. l unbounded-buffer places no practical limit on the size of the buffer. l bounded-buffer assumes that there is a fixed buffer size.

23 Bounded-Buffer – Shared-Memory Solution §Shared data var n; type item = … ; var buffer. array [0..n–1] of item; in, out: 0..n–1;

24 Bounded-Buffer – Shared-Memory Solution §Producer process repeat … produce an item in nextp … while in+1 mod n = out do no-op; buffer [in] :=nextp; in :=in+1 mod n; until false;

25 Bounded-Buffer (Cont.) §Consumer process repeat while in = out do no-op; nextc := buffer [out]; out := out+1 mod n; … consume the item in nextc … until false;

26 Bounded-Buffer (Cont.) §Solution is correct, but …?

27 Bounded-Buffer §Shared data #define BUFFER_SIZE 10 typedef struct {... } item; item buffer[BUFFER_SIZE]; int in = 0; int out = 0; int counter = 0;

28 Bounded-Buffer §Producer process item nextProduced; while (1) { while (counter == BUFFER_SIZE) ; /* do nothing */ buffer[in] = nextProduced; in = (in + 1) % BUFFER_SIZE; counter++; }

29 Bounded-Buffer §Consumer process item nextConsumed; while (1) { while (counter == 0) ; /* do nothing */ nextConsumed = buffer[out]; out = (out + 1) % BUFFER_SIZE; counter--; }

30 Bounded Buffer §The statements counter++; counter--; must be performed atomically. §Atomic operation means an operation that completes in its entirety without interruption.

31 Bounded Buffer §The statement “count++” may be implemented in machine language as: register1 = counter register1 = register1 + 1 counter = register1 §The statement “count - -” may be implemented as: register2 = counter register2 = register2 – 1 counter = register2

32 Process Synchronization §Cooperating processes executing in a system can affect each other §Concurrent access to shared data may result in data inconsistency §Maintaining data consistency requires mechanisms to ensure the orderly execution of cooperating processes

33 Race Condition §Problem in previous example: One process started using shared variable before another process was finished using it §Race condition: The situation where several processes access – and manipulate shared data concurrently. The final value of the shared data depends upon which process finishes last. §To prevent race conditions, concurrent processes must be synchronized.

34 Two processes want to access shared memory at the same time Debugging very difficult...

35 Example...

36 Solution #1

37 Solution #2

38 Solution #3

39 The Critical-Section Problem §n processes all competing to use some shared data §Each process has a code segment, called critical section, in which the shared data is accessed. §Problem – ensure that when one process is executing in its critical section, no other process is allowed to execute in its critical section.

40

41 §No processes may be simultaneously inside their critical sections §No assumptions may be made about the speeds or number of CPUs §No process running outside the critical region may block other processes §No process waits forever to enter its critical section

42 §Only 2 processes, P 0 and P 1 §General structure of process P i (other process P j ) do { entry section critical section exit section remainder section } while (1); §Processes may share some common variables to synchronize their actions.

43 Solution to Critical-Section Problem 1.Mutual Exclusion. If process P i is executing in its critical section, then no other processes can be executing in their critical sections.

44 Solution to Critical-Section Problem 2.Progress. If no process is executing in its critical section and some processes wish to enter their critical section, then only those processes that are not executing in the remainder section can participate in the decision on which will enter critical section next.

45 Solution to Critical-Section Problem 3.Bounded Waiting. A bound must exist on the number of times that other processes are allowed to enter their critical sections after a process has made a request to enter its critical section and before that request is granted. §Assume that each process executes at a nonzero speed §No assumption concerning relative speed of the n processes.

46 §Assume the basic machine language instructions like load or store etc. are executed atomically.

47 Algorithm 1 turn is initialized to 0 (or 1)