01 - ILP Flashcards

(30 cards)

1
Q

refers to an activity, item of information, material, or product, that is between the starting point and the completion point

A

pipeline

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

a pipeline is __ or __ in nature

A

sequential, linear

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

a pipeline is typical in __ and __

A

manufacturing, assembly lines

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

refers to an implementation technique where multiple instructions are overlapped in execution

A

pipelining

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

cons of not implementing pipelining (3)

A
  1. resources are wasted
  2. machines are not used at some point
  3. time is not optimized
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

refers to the three stages that compose a pipeline

A

fetch, decode, execute

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

true or false:

pipeline stages may vary and can have 5 stages, with 3 stages being the least

A

true, it varies depending on the design

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

refers to the degree to which, on average, the instructions of a program can be executed in parallel

A

instruction-level parallelism

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

refers to the stage that must be present in order to ensure correctness whenever stages do not properly align or data hazards are present

A

stall

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

on ILP:

we are concerned with maximizing the __ while keeping the correctness of the result

A

number of instructions that are executed in parallel

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

which of the following statement/s is/are true?

on the ideal pipeline:

a. all objects go through the same stages
b. sharing of resources is allowed between stages as long as they are not sequentially executed

A

a only, no sharing of resources between any two stages

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

__ through all pipeline stages is equal

A

propagation delay

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

scheduling of a transaction entering the pipeline is [affected/not affected] by the transactions in other stages

A

not affected

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

true or false:

the time consumed by a five stage pipeline is greater than the maximum between the time per stage (fetch, decode, execute, memory, write)

A

true

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

on the time consumed by a five-stage pipeline:

the max time consumed is typically the time spent on __, for most of the processors

A

memory

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

what does CPI stand for?

A

cycles per instruction

17
Q

CPI is [increased/decreased] when instructions are pipelined

18
Q

formula for time per program (time/program)

A

time/program = (instructions/program) * (cycles/instruction) * (time/cycle)

19
Q

CPI depends on __ and __

A

ISA, microarchitecture

20
Q

time/cycle depends on __ and __

A

microarchitecture, technology used

21
Q

refers to the execution of an action

22
Q

refers to work done based on a given set of resources

23
Q

performance, in pipelining, are typically measured through: (3)

A
  1. CPI
  2. average instruction execution time
  3. speedup
24
Q

formula for CPI

A

sum(numOfInstForInstrType * clkCycleForInstType) / totalNumberOfInsts

25
Suppose an N-segment pipeline processes M instructions without stalls or penalties. We know that it takes N-1 cycles to load (setup) the pipeline, and M cycles to complete the instructions . Thus, the number of cycles is given by:
Ncyc = N + M - 1
26
formula for CPI given Ncyc = N + M - 1
CPI = Ncyc / M CPI = 1 + ((N - 1) / M)
27
formula for number of cycles with stalls
Ncyc = N + M + K - 1
28
formula for CPI with stalls
CPI = 1 + (N + K - 1)/M
29
formula for the average instruction execution time of an unpipelined processor
Clock cycle * Average CPI
30
Formula for speedup from pipelining
AveInstrTimeUnpipelined / AveInstrTimePipelined