COMP 361 Lecture Notes - Lecture 1: Direct Memory Access, Device Driver, Multiprocessing
27 Apr 2018
School
Department
Course
Professor
COMP 361 – Unit 1 (Notes)
• Device Controller
o Maintains some local buffer storage and a set of special-purpose registers. (10)
▪ Responsible for moving the data between the peripheral devices that it controls and its local buffer
storage. (10)
o Typically, operating systems have a device driver for each device controller
• Device Driver
o Understands the device controller and presents a uniform interface to the device to the rest of the operating
system. (11)
• Interrupt-driven I/O Operation (11)
o Device driver loads the appropriate registers within the device controller.
o Device controller examines the contents of these registers to determine what action to take.
o Controller starts to transfer data from the device to its local buffer, and informs the device driver via interrupt
when the action is complete.
o The device driver then returns control to the OS, possibly returning the data or a pointer to the data if the
operation was a read.
o Produces high overhead costs when used with bulk data.
• Direct Memory Access (DMA) (11)
o After setting up buffers, pointers, and counters for the I/O device, the controller transfers an entire block of data
directly to or from its own buffer storage to memory, with no intervention by the CPU.
o Only one interrupt is generated per block, to tell the device driver that the operation is completed, rather than
one interrupt per byte.
• Single Processor Systems (12)
o If there is only one general-purpose CPU, then the system is a single-processor system.
▪ Devices can have an I/O processor, but this does not turn the computer into a multi-processor system.
• Multiprocessor System (12)
o Three main advantages:
▪ Increased Throughput: By increasing the number, we increase the throughput. This is not multiplied by
the number of processors, however, since a certain amount of overhead is incurred in keeping all the
parts working correctly.
▪ Economy of Scale: They can cost less, since they can share the same peripherals. (13)
▪ Increased Reliability: The failure of one CPU will not halt the system, only slow it down. (13)
• Graceful degradation is the ability to continue to provide proportional service to the level of
hardware.
• Fault Tolerance is when they can suffer a failure of a component and still continue to run.
o Asymmetrical multiprocessing
▪ Is when there is a master / slave processor relationship, where the master coordinates the actions of the
slaves. (13)
o Symmetrical multiprocessing (SMP)
▪ Many processors can run simultaneously, and have different workloads. One processor could sit idle,
while the other is overloaded.
• Clustered Systems (14)
o They are composed of two or more individual systems coupled together (LAN, or other fast connection).
o Provides high-availability, such that service will continue even if one ore more systems in the cluster fail.
o Adds a level of redundancy, such as concurrent monitoring that results adaptation to the failure of a single
machine.
o Asymmetrical Clustering is where one system monitors another in standby, and replaces it if it fails.
▪ Symmetrical mode is where both computers are running applications, and monitor each other.
o Distributed Locking Manager (DLM) is used to prevent simultaneous data access conflicts. (15)
• Multiprogramming (15)
o Increases CPU utilization by organizing jobs (code and data) so that the CPU always has an operation to
execute.
• Time-Sharing / Multitasking (16)
o The CPU executes multiple jobs by switching among them, but the switches occur so frequently that the users
can interact with each program while it is running.
▪ Requires an interactive (or hands-on) computer system, with a user directing the computer.
▪ A program loaded into memory is called a process.
▪ Since main memory is too small, jobs are kept in the job pool. ( 17)
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
If there is only one general-purpose cpu, then the system is a single-processor system: devices can have an i/o processor, but this does not turn the computer into a multi-processor system, multiprocessor system (12, three main advantages: Increased throughput: by increasing the number, we increase the throughput. Is when there is a master / slave processor relationship, where the master coordinates the actions of the slaves. (13: symmetrical multiprocessing (smp, many processors can run simultaneously, and have different workloads. If there are several jobs to be brought into main memory, and there is not enough space, then the system must choose between them. If several jobs are ready to be run at the same time, the system must choose between them, which is. Information is normally kept in some storage system (such as main memory), it is copied into a faster storage system (cache) on a temporary basis. (24)
