Operating system chapter 10
اسلاید 1: 1Multiprocessor and Real-Time SchedulingChapter 10
اسلاید 2: 2Classifications of Multiprocessor SystemsLoosely coupled or distributed multiprocessor, or clusterEach processor has its own memory and I/O channelsFunctionally specialized processorsSuch as I/O processorControlled by a master processorTightly coupled multiprocessingProcessors share main memory Controlled by operating system
اسلاید 3: 3Independent ParallelismSeparate application or jobNo synchronization among processesExample is time-sharing system
اسلاید 4: 4Coarse and Very Coarse-Grained ParallelismSynchronization among processes at a very gross levelGood for concurrent processes running on a multiprogrammed uniprocessorCan by supported on a multiprocessor with little change
اسلاید 5: 5Medium-Grained ParallelismSingle application is a collection of threadsThreads usually interact frequently
اسلاید 6: 6Fine-Grained ParallelismHighly parallel applicationsSpecialized and fragmented area
اسلاید 7: 7SchedulingAssignment of processes to processorsUse of multiprogramming on individual processorsActual dispatching of a process
اسلاید 8: 8Assignment of Processes to ProcessorsTreat processors as a pooled resource and assign process to processors on demandPermanently assign process to a processorKnown as group or gang schedulingDedicate short-term queue for each processorLess overheadProcessor could be idle while another processor has a backlog
اسلاید 9: 9Assignment of Processes to ProcessorsGlobal queueSchedule to any available processorMaster/slave architectureKey kernel functions always run on a particular processorMaster is responsible for schedulingSlave sends service request to the masterDisadvantagesFailure of master brings down whole systemMaster can become a performance bottleneck
اسلاید 10: 10Assignment of Processes to ProcessorsPeer architectureOperating system can execute on any processorEach processor does self-schedulingComplicates the operating systemMake sure two processors do not choose the same process
اسلاید 11: 11
اسلاید 12: 12Process SchedulingSingle queue for all processesMultiple queues are used for prioritiesAll queues feed to the common pool of processors
اسلاید 13: 13Thread SchedulingExecutes separate from the rest of the processAn application can be a set of threads that cooperate and execute concurrently in the same address spaceThreads running on separate processors yields a dramatic gain in performance
اسلاید 14: 14Multiprocessor Thread SchedulingLoad sharingProcesses are not assigned to a particular processorGang schedulingA set of related threads is scheduled to run on a set of processors at the same time
اسلاید 15: 15Multiprocessor Thread SchedulingDedicated processor assignmentThreads are assigned to a specific processorDynamic schedulingNumber of threads can be altered during course of execution
اسلاید 16: 16Load SharingLoad is distributed evenly across the processorsNo centralized scheduler requiredUse global queues
اسلاید 17: 17Disadvantages of Load SharingCentral queue needs mutual exclusionMay be a bottleneck when more than one processor looks for work at the same timePreemptive threads are unlikely resume execution on the same processorCache use is less efficientIf all threads are in the global queue, all threads of a program will not gain access to the processors at the same time
اسلاید 18: 18Gang SchedulingSimultaneous scheduling of threads that make up a single processUseful for applications where performance severely degrades when any part of the application is not runningThreads often need to synchronize with each other
اسلاید 19: 19Scheduling Groups
اسلاید 20: 20Dedicated Processor AssignmentWhen application is scheduled, its threads are assigned to a processorSome processors may be idleNo multiprogramming of processors
اسلاید 21: 21
اسلاید 22: 22Dynamic SchedulingNumber of threads in a process are altered dynamically by the applicationOperating system adjust the load to improve utilizationAssign idle processorsNew arrivals may be assigned to a processor that is used by a job currently using more than one processorHold request until processor is availableAssign processor a jog in the list that currently has no processors (i.e., to all waiting new arrivals)
اسلاید 23: 23Real-Time SystemsCorrectness of the system depends not only on the logical result of the computation but also on the time at which the results are producedTasks or processes attempt to control or react to events that take place in the outside worldThese events occur in “real time” and tasks must be able to keep up with them
اسلاید 24: 24Real-Time SystemsControl of laboratory experimentsProcess control in industrial plantsRoboticsAir traffic controlTelecommunicationsMilitary command and control systems
اسلاید 25: 25Characteristics of Real-Time Operating SystemsDeterministicOperations are performed at fixed, predetermined times or within predetermined time intervalsConcerned with how long the operating system delays before acknowledging an interrupt and there is sufficient capacity to handle all the requests within the required time
اسلاید 26: 26Characteristics of Real-Time Operating SystemsResponsivenessHow long, after acknowledgment, it takes the operating system to service the interruptIncludes amount of time to begin execution of the interruptIncludes the amount of time to perform the interruptEffect of interrupt nesting
اسلاید 27: 27Characteristics of Real-Time Operating SystemsUser controlUser specifies prioritySpecify pagingWhat processes must always reside in main memoryDisks algorithms to useRights of processes
اسلاید 28: 28Characteristics of Real-Time Operating SystemsReliabilityDegradation of performance may have catastrophic consequencesFail-soft operationAbility of a system to fail in such a way as to preserve as much capability and data as possibleStability
اسلاید 29: 29Features of Real-Time Operating SystemsFast process or thread switchSmall sizeAbility to respond to external interrupts quicklyMultitasking with interprocess communication tools such as semaphores, signals, and events
اسلاید 30: 30Features of Real-Time Operating SystemsUse of special sequential files that can accumulate data at a fast ratePreemptive scheduling base on priorityMinimization of intervals during which interrupts are disabledDelay tasks for fixed amount of timeSpecial alarms and timeouts
اسلاید 31: 31Scheduling of a Real-Time Process
اسلاید 32: 32Scheduling of a Real-Time Process
اسلاید 33: 33Real-Time SchedulingStatic table-drivenDetermines at run time when a task begins executionStatic priority-driven preemptiveTraditional priority-driven scheduler is usedDynamic planning-basedFeasibility determined at run timeDynamic best effortNo feasibility analysis is performed
اسلاید 34: 34Deadline SchedulingReal-time applications are not concerned with speed but with completing tasks
اسلاید 35: 35Deadline SchedulingInformation usedReady timeStarting deadlineCompletion deadlineProcessing timeResource requirementsPrioritySubtask scheduler
اسلاید 36: 36Two Tasks
اسلاید 37: 37
اسلاید 38: 38
اسلاید 39: 39
اسلاید 40: 40Rate Monotonic SchedulingAssigns priorities to tasks on the basis of their periodsHighest-priority task is the one with the shortest period
اسلاید 41: 41Periodic Task Timing Diagram
اسلاید 42: 42
اسلاید 43: 43Priority InversionCan occur in any priority-based preemptive scheduling schemeOccurs when circumstances within the system force a higher priority task to wait for a lower priority task
اسلاید 44: 44Unbounded Priority InversionDuration of a priority inversion depends on unpredictable actions of other unrelated tasks
اسلاید 45: 45Priority InheritanceLower-priority task inherits the priority of any higher priority task pending on a resource they share
اسلاید 46: 46Linux SchedulingScheduling classesSCHED_FIFO: First-in-first-out real-time threadsSCHED_RR: Round-robin real-time threadsSCHED_OTHER: Other, non-real-time threadsWithin each class multiple priorities may be used
اسلاید 47: 47
اسلاید 48: 48Non-Real-Time SchedulingLinux 2.6 uses a new scheduler the O(1) schedulerTime to select the appropriate process and assign it to a processor is constantRegardless of the load on the system or number of processors
اسلاید 49: 49
اسلاید 50: 50UNIX SVR4 SchedulingHighest preference to real-time processesNext-highest to kernel-mode processesLowest preference to other user-mode processes
اسلاید 51: 51UNIX SVR4 SchedulingPreemptable static priority schedulerIntroduction of a set of 160 priority levels divided into three priority classesInsertion of preemption points
اسلاید 52: 52SVR4 Priority Classes
اسلاید 53: 53SVR4 Priority ClassesReal time (159 – 100)Guaranteed to be selected to run before any kernel or time-sharing processCan preempt kernel and user processesKernel (99 – 60)Guaranteed to be selected to run before any time-sharing processTime-shared (59-0)Lowest-priority
اسلاید 54: 54SVR4 Dispatch Queues
اسلاید 55: 55Windows SchedulingPriorities organized into two bands or classesReal timeVariablePriority-driven preemptive scheduler
اسلاید 56: 56
اسلاید 57: 57
نقد و بررسی ها
هیچ نظری برای این پاورپوینت نوشته نشده است.