Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
87 Cards in this Set
- Front
- Back
Infiniband allows IO devices to be placed outside the server chassis, in some cases as much as 6 miles away
|
true
|
|
The DMA mechanisms can be configured in different ways, including
|
-There is a single bus. The DMA module and each IO module are attached to the Bus
-Only the DMA module is attached to the bus. The IO modules are not attached to the bus, but rather, are attached to the DMA module -The IO modules are attached to an IO bus with is attached to the DMA module which is attached to the system bus |
|
Reasons a peripheral device is not connected directly to the system bus, but rather to the system bus through an IO module include:
|
-removes burden of logic from CPU
-devices are slower that mem or CPU -devices can have high bandwidth requirements -defferent peripherals can have different formats |
|
When using the DMA IO technique, there is a DMA module instead of an IO module
|
False
|
|
What are the 2 possible mode of addressing when the CPU, mem, and IO module share a common bus?
|
-memory mapped
-isolated |
|
_____is the type of IO in which a program issues an IO request, and then continues to execute other instructions until it is informed by the IO hardware that the io operation is complete
|
interrupt driven IO
|
|
with isolated IO, there is an address space for mem and a separate address for devices. It is not the value of the address that determines if an address is a momory localtion or a device, rather a command line on the bus indicates whether the address is for a mem location or a device
|
true
|
|
vector means that a device places a unique ID on the data lines
|
true
|
|
what is the technique which IDs the interrupting device by means of a shared interrupt line?
|
daisy chain
|
|
usually external devices are connected to the computer via an IO module
|
true
|
|
what technique IDs the interrupting device by the following method. the IO module which wants to issue an interrupt gains control of the bus. when the CPU acks the interrupt, the IO module places its vector in the data lines
|
bus arbitration
|
|
infiniband is an IO spec aimed at the high end server market
|
true
|
|
firewire provides for auto config so that it isnt neccesary to mannually set device IDs, or to be concerned with the relative position of devices
|
true
|
|
on of the motivations for firewire is to provide a single IO interface with a simple connector that can handle numerous devices through a single port
|
true
|
|
in trying to ID which device issued an interrupt, one strategy is to use multiple interrupt lines and each device has its own line
|
false
|
|
isochronous transmission requires an ack from the reciever
|
false
|
|
when compenents operate at different speeds, a ____ is used to queue up data
|
buffer
|
|
an IO channel which controls multiple devices and can transfer data with several of these devices at a time is a
|
multiplexor channel
|
|
the computers IO subsystem us uts interface to the outside world
|
true
|
|
which is the common name for the IEEE standard for a high performance serial bus
|
firewire
|
|
___ is the type of IO in which the IO occurs under the direct and continous control of the program that is requesting the IO operation
|
programmed I/o
|
|
what is it called when the DMA module forces the CPU to suspend operation temporarily so that the DMA module can gain control of the bus?
|
cycle stealing
|
|
when multiple devices issue an interrupt, there must be some sort of prioritizing to ensure that the most important ones are handled first. which of the following is a technique for prioritizing?
|
-multiple lines and each line has a priority
-in order devices are polled -bus arbitration can employ a scheme |
|
firewire and infiniband are examples of external IO interfaces
|
true
|
|
in interrupt driven IO, since there can be multiple devices on each IO module, and since there can be multiple IO modules, how does the CPU determine which device issued the interrupt?
|
-use multiple interrupt lines
-software polling -daisy chain -vectored bus arbitration |
|
when there is an IO request, the CPU is responsible for moving data between mem and the IO module
|
false
|
|
an IO module interfaces between the computer and one or more peripheral devices. the IO module may vary in complexity. what is the lowest level IO module called?
|
i/o controller
|
|
IO controllers are common on mainframes
|
false
|
|
mem and peripherals operate at differnt speeds. the io module must be able to operate at both speeds
|
true
|
|
in an IO module, the ___ converts data from electrical to other forms of energy
|
transducer
|
|
ID the principle IO techniques
|
-programmed io
-interrupt driven io -DMA |
|
an io channel which controls multiple hi speed devices and is dedicated to the transfer of data is called a
|
selector channel
|
|
isochronous and asynchronous transmissions can happen on the same line
|
true
|
|
the major disadvantage of ptogrammed io is that it keeps the CPU busy
|
true
|
|
if there are 8 bits for addressing, in memory mapped IO that is the total number of locations?
|
256
|
|
an IO module is which of the following
|
-interfaces to the system bus
-controls one or more peripherals -physically connects to the comm medium -contains logic for performing comm between bus and peripheral |
|
which of the following IO requires the most of the CPU
|
programmed io
|
|
which of the following are characteristics of firewire?
|
-hi speed
-low cost -easy to implement -used in consumer market -serial transmission -hot plugging |
|
_is the type of io in which a specialized IO proccessor takes over control of an IO operation to move a large block of data
|
DMA
|
|
how many locations can be mapped in isolated IO with 8 bits?
|
512
|
|
the node responsible for coordinating the switch between asynchronous and isochronous transmission is the
|
cycle master
|
|
interrupt driven io is more efficient, but every byte transferred must be passed through the CPU
|
true
|
|
fair arbitration divides time into fairness intervals. once a device has the bus, it cannot request the bus diring the same fairness intervak
|
true
|
|
the transmission mode that provides a variable amount of data to be transferred in a sequence of fixed size packets is called
|
isochronous
|
|
each IO device has an address
|
true
|
|
software polling is more efficiant that daisy chaining
|
false
|
|
which IO techniques requires the least CPU
|
DMA
|
|
firewire was developed as a low cost alternative to more expensive io channels for large computers
|
true
|
|
io channels are common on microcomputers
|
false
|
|
an io module performs error detection with hamming code
|
false
|
|
infiniband is designed for comm between cpu and intelligent devices and is intended to replace pci
|
true
|
|
by daisy chaining, firewire can connect up to 200 devices
|
false
|
|
the ability to plug and unplug devices without powering down is called
|
hot plugging
|
|
with mem mapped io, there is a single address space for mem locations and io devices.
|
true
|
|
when trying to ID which device has issued an interrupt, what is the technique ised to query each IO module to find out which module issued the interrupt?
|
software polling
|
|
dma is the most appropriate io technique when large volumes of data need to be moves
|
true
|
|
a multiplexor channel dedicated to hi speed devices is called a
|
block multiplexor
|
|
firewire and infiniband are important point to point interfaces
|
false
|
|
firewire can support more peripherals than can practically be needed since as many as 1022 buses can be interconnected
|
true
|
|
a multiplexor channel dedicated to slow speed devices is called a
|
byte multiplexor
|
|
the disadvantage of software polling is that it is time consuming
|
true
|
|
an io module interfaces between the computer and one ore more peripheral devices. the most high level io module is called a
|
i/o channel
|
|
since the bottlenecks in a computer is the IO, parralell interfaces are becoming more common
|
false
|
|
_____allows for effective multiprogramming. it creates the illusion that there is more memory than there really is
|
virtual memory
|
|
a special cache devoted specifically to storing page frams is called a
|
transition lookaside buffer
|
|
reasons a program may cause an interrupt to occur include
|
-program produced an error
-timeout |
|
with variable sized partitions
|
-the exact amount of mem required can be allocated
-more efficient that fixed -can still waste space due to fragmentation |
|
the data structure an os creates to store info about a process is called a
|
process control block
|
|
when using fixed partitions, each partition is the same size
|
false
|
|
the CPU can cease to execute a program and begin executing instructions from the OS ofr the folloowing reasons
|
-program issues a request to the os
-program causes an interrupt -another interrupt happens |
|
when a job is admitted for processing, it becomes a process
|
true
|
|
the key to multiprogramming is scheduling
|
true
|
|
the mom management technique of dividing mem into small equal sized partitions is called
|
paging
|
|
what is it called when a program references an instruction or data not in mem?
|
page fault
|
|
which type of scheduling deals with jobs that are swapped out?
|
medium term
|
|
in paging, a process can be allocated potentially many non contigous page frams. a page table is maintained for each process to keep track of page frams
|
true
|
|
with demand paging, the technique of replaceing an idle process that is not in the ready queue with another process is called
|
swapping
|
|
a technique for overcoming wasted space due to fragmentation by compressing holes is
|
compaction
|
|
relative to other types of scheduling, the short term scheduler executes frequently
|
true
|
|
simplest scheme for partitioning is fixed size
|
true
|
|
the task of subdiviing mem to accomidate multiple process is called
|
memory management
|
|
the long term scheduler determines which programs are addmitted to the ready queue and thus controls the degree of multi programming
|
true
|
|
the technique of bringing in a page of a process only when needed is called
|
demand paging
|
|
the goal of mem management is to keep the CPU busy
|
true
|
|
short term scheduler is also known as
|
dispatcher
|
|
the long term scheduler executes very frequently
|
false
|
|
segmentation is another was in which addressable mem can be divided. unlike paging, segmentation is visible to the programms. segmentation alllwos the programmer to view memory as consisting of multiple address spaces or segments
|
true
|