Week 6 Flashcards

1
Q

What is Parallel Computing?

A

Improve performance by performing multiple computations in parallel

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

What is Flynn’s Taxonomy?

A

Classification of computer architectures based on number of instruction and data streams they can handle simultaenously

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

What is an Instruction Stream?

A

Sequence of instructions; arithmetic, data transfer and branching

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

What is a Data Stream?

A

Consists of operands (constants, variables and memory addresses)

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

What is Single Instruction Single Data (SISD)?

A

Sequential computer with no data or instruction parallelism

Older single-core CPUs and simple microcontrollers

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What is Multiple Instruction, Single Data (MISD)?

A

Multiple instructions operate on one piece of data simultaneously

Used for fault tolerance

Rare in practice

Eg. Space shuttle flight systems

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

What is Triple Modular Redundancy?

A

Majority vote of three independent systems or processors

Space satellite systems often use TMR

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

What is Single Instruction, Multiple Data (SIMD)?

A

Execute the same instruction on multiple data points simultaneously

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

What are SIMD computers good for?

A

Good for vector operations including image and audio processing

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

What are SIMD computers not good for?

A

SIMD computers are not suitable when the type of instruction depends on the data

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

What is a GPU (Graphics Processing Unit)?

A

GPU contains many ALUs to perform mathematical operations in parallel

GPU based on SIMD architecture

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

How to GPUs compare to CPUs?

A

GPUs have more ALUs and rely less on cache

GPUs have higher throughput and higher latency

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

What are Threads?

A

GPUs hide latency by utilising threads

Process may be subdivided into multiple threads

Threads within a process share data, heap, and code, but have their own registers and stack

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

What is Multiple Instruction, Multiple Data (MIMD)?

A

Multiple processors simultaneously execute different instructions on different pieces of data

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

What are Multi Core Processors?

A

A multi-core processor puts many CPU cores on one chip

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

What are Multiprocessors?

A

Two or more CPUs within a single computer system

CPUs communicate through shared memory variables

To modify shared variable, CPU locks variable so that no other processor can modify it

17
Q

What are Multicomputers?

A

Different computers connected together

18
Q

What is the benefit of Shared Memory (Multiprocessors)?

A

Faster communication

19
Q

What is the benefit of Distributed Memory (Multicomputers)?

A

Allows for higher bandwidth

20
Q

What are the 2 energy problems the world faces?

A
  1. Energy produces greenhouse gas and emission
  2. Energy Poverty - low emissions of the poor as they lack access to modern energy or tech
21
Q

What is the solution to the 2 energy problems?

A
  1. Reducing greenhouse gas emissions
  2. Removing gases from the atmosphere
22
Q

What is the Energy Demand for ICT?

A

ICT energy demand might rise to 8% of total electricity demand by 20230, in the best case, or 21% in the worst case

ICT sector used an estimated 4-6% of all electricity generated globally in 2020

23
Q

What is the Quick and the Dead?

A

Historically (like the Wild West), the processor world consisted of the quick (those with high SPEC marks) and the dead (those without)

24
Q

What is the Energy-Efficient and the Dead?

A

Today, the processor world increasingly consists of the energy-efficient (those with high SPEC marks/watt) and the dead (those without)

25
What is the Sleep State?
Up until circa 1990 there were only 2 power states On Off
26
What are Intermediary states (like sleep mode)?
Low power consumption Quick recovery to active state based on user input
27
What is the Advanced Configuration and Power Interface (ACPI)?
A standard device config and power management by the OS Allows OS to control the amount of power supplied to various devices using different power states
28
What are Global States G0-G3?
Apply to the entire system and are visible to the user
29
What are Device States D0-D3?
States of particular devices (eg. disks, keyboards) generally not visible to the user
30
What are Processor States C0-C3?
Processor power consumption and thermal management states - substates of G0 Processor states applied to individual cores Synergy effects when multiple cores within same processor are shut down Smartphones have both high-performance cores and energy saving cores
31
What are Performance States P0-P15?
Power consumption and capability states within the active states (substates of D0/C0) Processor could save energy by scaling down power and clock frequency
32
What is Power Usage Effectiveness?
PUE is a metric used to determine energy efficiency of a data centre PUE = Total facility energy/ IT energy = 1 + Non IT facility energy/ IT equipment energy Lower the PUE more efficient
33
What are the values of the PUE chart?
PUE Level of efficiency 1.0 Ideal 1.2 Very efficient 2 Average 2.5 Inefficient 3 Very Inefficent
34
What is the Open Compute Project (OCP)?
OCP is a collaborative community focused on redesigning hardware to efficiently support the growing demands on compute infrastructure Best practices for improving energy efficiency, cost and scalability
35
What are 3 ways to Lower Power Usage Effectiveness
Bespoke Hardware : Decide what to include and remove (Do you need GPUs, lights)? Smarter Cooling: Use locations with a cold climate Mobile Computing: Let mobile devices do a share of the work