Upload
kristin-nichols
View
218
Download
0
Embed Size (px)
Citation preview
Evaluation of Windows 7 RC Build 7100
By
Muswera Walter
Supervisor: Mr John Ebden
Consultants: Billy Morgan and Jill Japp
Background
• Vista was not received warmly due to bugs, crashes, poor performance and software incompatibility and demand for system resources
• Microsoft has been working on a new Operating System (OS)
• Available release at time of evaluation was Windows 7 RC Build 7100 (May 2009)
Motivation
Claims have been made about improvements that have been made on Windows 7 since the XP era and the ill fated Vista era: Security
Software compatibility
Enhanced performance (graphics, disk I/O, robustness etc)
Ease and speed of installation
Fast start-up and shutdown time
User friendliness
This assessment was carried out to validate some of these claims
Importance of this Research
Every new OS should be tested/evaluated before being deployed at full scale to allow for:
Planning of transition (upgrade and training)
Avert loss of time and revenue through crashes and incompatibilities
Evaluate availability of system support
Features Evaluated
• Security
• Virtualisation
• Software compatibility
• Performance benchmarking
Specifications of Experimental Machine
• Intel Core 2 Quad processor (Q9400 - 2.66GHz, 6MB L2 cache)
• 4GB RAM
• 500GB SATA hard drive with 32MB cache
• 256MB graphics with DirectX 10 support
• DVI and VGA connector
• DVD writer with LightScribe
• G45M Digital motherboard with 1GB network port
Minimum System Requirements for Windows 7
• 1 GHz 32-bit or 64-bit processor
• 1 GB of system memory
• 16 GB of available disk space
• support for DirectX 9 graphics with 128 MB memory (to enable the Aero theme)
• DVD-R/W Drive
• Internet access for updates
Same as Vista
Testing Conditions• The same computer hardware was used for all tests
• Both OS were installed fresh in a dual boot setup
• Results obtained reflect what you get when you install a plain OS without any tweaks or configuration changes (standard install)
• Install time was measured from the moment the option to boot from CD/DVD was chosen until the point where the desktop reached.
• Windows 7 Ultimate (32bit RC Build 7100)
• XP Professional (32bit SP3)
Security
• Bitlocker and Bitlocker To Go 100MB partition TPM and PCR Usability on other PCs running other Oses Bitlocker To Go Reader
• UAC Implementation of the 4 new levels Is it really secure?
Virtualisation
• Windows XP Mode
• consists of two things:
the Windows Virtual PC engine
a licensed copy of Windows XP SP3
• System requirements
2GB of memory
15GB disk space/virtual machine
Software Compatibility tests
• Successful Installations
• Compatibility issues
Drivers (TPM)
Nero 7 (program has known compatibility issues)
Daemon Tools (program blocked due to compatibility
issues)
Tests carried out on XP SP3 and Windows 7
• Performance benchmarking
• Installation time for each operating system
• Logon, boot up and shutdown time
Performance
• Passmark PerformanceTest version 7.0 was used to assess perfomance
CPU test suite is designed to measure CPU operations
Memory test suite is designed to exercise the memory sub-system of the computer.
Disk test suite is designed to measure the rate at which files are read and written from and to disk
2D Graphics test suite tests of standard two dimensional graphical functions
3D Graphics test suite measures the performance of the 3D graphics
CPU Tests
0.00
1000.00
2000.00
3000.00
4000.00
5000.00
6000.00
7000.00
516.
98
1982
.40
1095
.22
5984
.44
229.
30
2930
.10
518.
04
1979
.82
1101
.22
6062
.06
230.
52
2955
.70
Windows 7 Windows XP Units of Measurement
• Integer - Millions of operations per second
• Floating Point - Millions of operations per second
• Find Prime numbers Thousands of Primes per second
Compression - Kilobytes Processed per second
Physics - Frames per second
Random String Sorting - Thousand Strings per second
Memory Tests
0.00
500.00
1000.00
1500.00
2000.00
2500.00
3000.00
2077
.50
1907
.24
1924
.24
2547
.76
910.
48
2077
.64
1924
.14
1961
.10
2727
.20
1061
.90
Windows 7 Windows XP
Units of Measurement
• Cached - Megabytes transferred per second
• Uncached - Megabytes transferred per second
• Write - Megabytes transferred per second
• Allocate small block - Megabytes transferred per second
• Large RAM - Operations per second
Disk Tests
Sequential Read Sequential Write Random Seek + RW0
20
40
60
80
100
120
93
81.2
3.22
97.14
84.9
3.25
Windows 7 Windows XP Units of Measurement
• Disk Sequential Write - Megabytes transferred per second
• Disk Sequential Read - Megabytes transferred per second
• Disk Random Seek + RW - Megabytes transferred per second
2D Graphics Tests
0.00
50.00
100.00
150.00
200.00
250.00
300.00
350.00
400.00
450.00
1.36
1.33
91.1
6 112.
64
66.2
8
386.
12
278.
78
8.98
2.98
124.
80
109.
18 137.
86
386.
82
274.
56
Windows 7 Windows XP Units of Measurement
• Solid Vectors - Thousands of vectors drawn per second
• Transparent Vectors - Thousands of vectors drawn per second
• Complex Vectors - Complex vectors drawn per second
• Fonts and Text - Operations per second
• Windows Interface Operations per second
• Image Filters - Filters per second
• Image Rendering - Images per second
3D Graphics Tests
Simple Medium Complex0.00
50.00
100.00
150.00
200.00
250.00
300.00
350.00
400.00
450.00
394.18
82.18
14.74
376.62
80.72
15.00
Windows 7 Windows XPUnits of Measurement
•Simple - Frames per second
•Medium - Frames per second
•Complex - Frames per second.
•DirectX 11 issues on complex graphics test
Summary of Findings• BitLocker Drive Encryption and Bitlocker To Go in Windows 7
require special hardware
• XP Mode also requires special hardware
• BitLocker Drive Encryption and Bitlocker To Go are only available in the Ultimate and Enterprise Editions
• XP Mode is available only in the Professional, Ultimate and Enterprise Editions
• Windows 7 is incompatible with some software
• Windows XP dominated most of the benchmarks run
• Windows 7 performs better in the 3D graphics
Conclusions
• XP still remains a better desktop OS since:
– it has few application compatibility issues
– works on ordinary hardware and
– has a better performance.
• disadvantage of “black box” benchmarking is that it cannot conclusively explain all of the performance differences– it merely exposes the differences