Defining Performance: Which Airplane Has The Best Performance?

Morgan Kaufmann Publishers August 20, 2021
Defining Performance
 Which airplane has the best performance?
Boeing 777 Boeing 777
BAC/Sud BAC/Sud
Concorde Concorde
Douglas Douglas DC-
DC-8-50 8-50
0 100 200 300 400 500 0 2000 4000 6000 8000 10000
Passenger Capacit y Cruising Range (miles)
BAC/Sud BAC/Sud
Concorde Concorde
Douglas Douglas DC-
DC-8-50 8-50
0 500 1000 1500 0 100000 200000 300000 400000
Cruising Speed (mph) Passengers x mph
Basic Metrics
Chapter 1 — Computer Abstractions and Technology 1

Response Time and Throughput
• throughput: work per unit time

– = (1 / latency) when there is NO OVERLAP
10 time units
– > (1 / latency) when there is overlap Finish
each
• in real processors there is always overlap time unit
– good metric for fixed amount of time (maximize work)
Throughput vs. Latency

• What is better?
– A machine that always takes 1 ns to do “task X” 1 time
– A machine that takes 15 ns to do “task X” 30 times...
• ...but 5 ns to do “task X” 1 time
– Machine 1:
• a lower latency for a single operation...
– Machine 2:
• better throughput for multiple operations
– What’s better?
• depends on what kind of computation you need to do

Where latency matters
Ratios of Measurements

Relative Performance
 Define Performance = 1/Execution Time
 “X is n time faster than Y”
Performanc e X Performanc e Y
 Execution time Y Execution time X  n
 Example: time taken to run a program

 10s on A, 15s on B
 Execution TimeB / Execution TimeA
= 15s / 10s = 1.5
 So A is 1.5 times faster than B
Example: Speedup

Speedup and % Increase and Decrease
Mean (Average) Performance Numbers
10

For Example…
11
Average execution time

Computer X (s) Computer Y (s) Speedup (X to Y)
App A 9 18 2
App B 10 7 0.7 1.63
App C 5 11 2.2
Ave. execution 8 12 1.5
time
7.66 11.15 1.456
12

Mean (Average) Performance

Numbers
13
Derived metrics
 Often we care about multiple metrics at once.
 Examples (Bigger is better)
 Bandwidth per dollar (e.g., in networking (GB/s)/$)
 BW/Watt (e.g., in memory systems (GB/s)/W)
 Work/Joule (e.g., instructions/joule)
 In general: Multiply by big-is-better metrics, divide by
smaller-is-better
 Examples (Smaller is better)
 Cycles/Instruction (i.e., Time per work)
 Latency * Energy --“Energy Delay Product”
 In general: Multiply by smaller-is-better metrics, divide
by bigger-is-better
14

Example: Energy-Delay
15
Example: Energy-Delay
16

What’s the Right Metric?
17
Measuring Execution Time

 Elapsed time
 Total response time, including all aspects
 Processing, I/O, OS overhead, idle time
 Determines system performance
 CPU time
 Time spent processing a given job
 Discounts I/O time, other jobs’ shares
 Comprises user CPU time and system CPU
time
 Different programs are affected differently by
CPU and system performance

CPU Clocking
 Operation of digital hardware governed by a
constant-rate clock
Clock period
Clock (cycles)
Data transfer
and computation
Update state
 Clock period: duration of a clock cycle

 e.g., 250ps = 0.25ns = 250×10–12s
 Clock frequency (rate): cycles per second
 e.g., 4.0GHz = 4000MHz = 4.0×109Hz

CPU Time
CPU Time  CPU Clock Cycles  Clock Cycle Time
CPU Clock Cycles

Clock Rate
 Performance improved by
 Reducing number of clock cycles (cycle count)
 Increasing clock rate
 Hardware designer must often trade off clock
rate against cycle count
CPU Time Example

 Computer A: 2GHz clock, 10s CPU time
 Designing Computer B
 Aim for 6s CPU time
 Can do faster clock, but causes 1.2 × clock cycles
 How fast must Computer B clock be?
Clock CyclesB 1.2  Clock Cycles A
Clock RateB  
CPU Time B 6s
Clock Cycles A  CPU Time A  Clock Rate A
 10s  2GHz  20  10 9
1.2  20  10 9 24  10 9
Clock RateB    4GHz
6s 6s

Instruction Count and CPI

Clock Cycles  Instructio n Count  Cycles per Instructio n
CPU Time  Instructio n Count  CPI  Clock Cycle Time
Instructio n Count  CPI

Clock Rate
 Instruction Count for a program
 Determined by program, ISA and compiler
 Average cycles per instruction (CPI)
 Determined by CPU hardware (e.g. multi-issue)
 If different instructions have different CPI
 Average CPI affected by instruction mix
CPI Example
 Computer A: Cycle Time = 250ps, CPI = 2.0
 Computer B: Cycle Time = 500ps, CPI = 1.2
 Same ISA
 Which is faster, and by how much?
CPU Time  Instruction Count  CPI  Cycle Time
A A A
 I  2.0  250ps  I  500ps A is faster…
CPU Time  Instruction Count  CPI  Cycle Time
B B B
 I  1.2  500ps  I  600ps
CPU Time
B  I  600ps  1.2
…by this much
CPU Time I  500ps
A

CPI in More Detail

 If different instruction classes take different
numbers of cycles
n
Clock Cycles   (CPIi  Instruction Count i )
i1
 Weighted average CPI

Clock Cycles n
 Instruction Count i 
CPI     CPIi  
Instruction Count i1  Instruction Count 
Relative frequency
CPI Example
 Alternative compiled code sequences using
instructions in classes A, B, C
Class A B C
CPI for class 1 2 3
IC in sequence 1 2 1 2
IC in sequence 2 4 1 1
 Sequence 1: IC = 5  Sequence 2: IC = 6
 Clock Cycles  Clock Cycles
= 2 ×1 + 1 ×2 + 2 ×3 = 4×1 + 1×2 + 1×3
= 10 =9
 Avg. CPI = 10/5 = 2.0  Avg. CPI = 9/6 = 1.5

CPI Example
27
Measuring CPI
28

The Iron Law of Performance
The Iron Law of Performance

Reducing Cycle Time
31
Example 1: Reduce the IC
32

Other Impacts on Instruction Count
33
Cycles Per Instruction
34

Example: Reducing CPI
35
Example: Reducing CPI
36

Performance Summary
 Performance depends on
 Algorithm: affects IC, possibly CPI
 More multiplications, higher CPI
 Programming language: affects IC, CPI
 Statements in the language are translated into
processor instructions
 Compiler: affects IC, CPI
 Compiler selects the instructions
 Use registers to eliminate loads and stores
 Instruction set architecture: affects IC, CPI,
clock rate
Pitfall: Amdahl’s Law

• Qualifies performance gain
• Amdahl’s Law defined…

– The performance improvement to be gained from using
some faster mode of execution is limited by the amount of
time the enhancement is actually used.
• Amdahl’s Law defines speedup:

Execution time for entire task without enhancement
Speedup =
Execution time for entire task using enhancement
when possible

Amdahl’s Law in Action
39
Amdahl’s Law in Action
40

Amdahl’s Law
41
End here
42

Amdahl’s Law
43
Amdahl’s Law
44

Amdahl’s Law
45
Amdahl’s Corollary
46

Defining Performance: Which Airplane Has The Best Performance?

Uploaded by

Copyright:

Available Formats

You might also like

Defining Performance: Which Airplane Has The Best Performance?

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Defining Performance: Which Airplane Has The Best Performance?

Uploaded by

Copyright:

Available Formats

Morgan Kaufmann Publishers August 20, 2021

Boeing 777 Boeing 777

Boeing 747 Boeing 747

Passenger Capacit y Cruising Range (miles)

Boeing 777 Boeing 777

Boeing 747 Boeing 747

0 500 1000 1500 0 100000 200000 300000 400000

Cruising Speed (mph) Passengers x mph

Chapter 1 — Computer Abstractions and Technology 1

Response Time and Throughput

• throughput: work per unit time

– good metric for fixed amount of time (maximize work)

Throughput vs. Latency

Chapter 1 — Computer Abstractions and Technology 2

Where latency matters

Chapter 1 — Computer Abstractions and Technology 3

 Example: time taken to run a program

Chapter 1 — Computer Abstractions and Technology 4

Speedup and % Increase and Decrease

Mean (Average) Performance Numbers

Chapter 1 — Computer Abstractions and Technology 5

Average execution time

Chapter 1 — Computer Abstractions and Technology 6

Mean (Average) Performance

Chapter 1 — Computer Abstractions and Technology 7

Chapter 1 — Computer Abstractions and Technology 8

What’s the Right Metric?

Measuring Execution Time

Chapter 1 — Computer Abstractions and Technology 9

 Clock period: duration of a clock cycle

Chapter 1 — Computer Abstractions and Technology 10

CPU Time Example

Chapter 1 — Computer Abstractions and Technology 11

Instruction Count and CPI

Chapter 1 — Computer Abstractions and Technology 12

CPI in More Detail

 Weighted average CPI

Chapter 1 — Computer Abstractions and Technology 13

Chapter 1 — Computer Abstractions and Technology 14

The Iron Law of Performance

The Iron Law of Performance

Chapter 1 — Computer Abstractions and Technology 15

Reducing Cycle Time

Example 1: Reduce the IC

Chapter 1 — Computer Abstractions and Technology 16

Other Impacts on Instruction Count

Cycles Per Instruction

Chapter 1 — Computer Abstractions and Technology 17

Example: Reducing CPI

Example: Reducing CPI

Chapter 1 — Computer Abstractions and Technology 18

Pitfall: Amdahl’s Law

• Amdahl’s Law defined…

• Amdahl’s Law defines speedup:

Chapter 1 — Computer Abstractions and Technology 19

Amdahl’s Law in Action

Amdahl’s Law in Action