Understanding and Using PC Bench (TM) 8.0

Copyright (c) 1992, 1993 by Ziff-Davis Publishing Company, L.P.
All rights reserved.
Manual release date: November 1993 with PC Bench version 8.0

LICENSE AGREEMENT FOR ZIFF-DAVIS' PC BENCH (TM) 8.0 

READ THIS AGREEMENT CAREFULLY BEFORE USING THE SOFTWARE EMBODIED IN THE 
PC BENCH{SYMBOL 228 \f "Symbol"} 8.0 DISKETTE (OR, IF DOWNLOADED, IN THE 
DOWNLOADED FILE(S)). Embodied in the PC Bench 8.0 diskette ("diskette") (or, 
if downloaded, in the downloaded file(s)) is the PC Bench 8.0 computer 
program and related documentation (the "Software"). Ziff-Davis Publishing 
Company, L.P., having a place of business at One Park Avenue, New York, New 
York 10016 ("Ziff") is the licensor under this Agreement and you are the 
licensee. By using the Software, in whole or in part, you agree to be bound 
by the terms of this Agreement. If you do not agree to the terms of this 
Agreement, promptly return the Software (or, if downloaded, delete the 
Software) to the Ziff-Davis Benchmark Operation at One Copley Parkway, Suite 
510, Morrisville, North Carolina 27560. Title to the Software and all 
copyrights, trade secrets and other proprietary rights therein are owned by 
Ziff. All rights therein, except those expressly granted to you in this 
Agreement, are reserved by Ziff.

1.      Limited License

This Agreement grants you only limited rights to use the Software. Ziff 
grants you a non-exclusive, non-transferable license to use the Software on a 
single dedicated computer or on a file server networked with multiple PC 
computers for the sole purpose of conducting benchmark tests to measure the 
performance of computer hardware and operating system configurations. You 
have the right to make a single copy of the Software for archival purposes 
and the right to transfer a copy of the Software across a network only to the 
PC computers attached to the network. Ziff hereby grants you the right to 
publish benchmark test results obtained by you from your use of the Software 
provided that with the publication of each such result you: 

A.      Identify Ziff-Davis, the name and version number of the benchmark 
	Software used (i.e., Ziff-Davis' PC Bench (TM) Ver. 8.0);

B.      Identify the exact name, processor speed and type, amount of RAM, 
	amount of secondary RAM cache, if any, hard disk size, type of hard 
	disk controller, and size of hardware hard disk cache, if any, of the 
	PC used for the test (e.g., WXY Corp. Model 466 with 66-MHz Intel (R) 
	486DX2-66 CPU, 8 MB of RAM, 64 KB RAM cache, 200 MB hard disk, IDE 
	controller, and no hardware disk cache); 

C.      Identify the exact video adapter name and refresh rate that produced 
	the result (e.g., XYZ Corp. XYZ Video adapter with a refresh rate of 
	72 Hz); 

D.      Identify the operating system version (e.g., MS-DOS (R) 5.0), size 
	and type of software disk cache, if any (e.g., SMARTDRV 2MB cache), 
	video resolution and color depth (e.g., 800 by 600 pixels with 256 
	colors), and any other special conditions used to achieve the result 
	(e.g., disk compression utility ABC version 1.0 enabled); 

E.      State that all products used in the test were shipping versions 
	available to the general public;

F.      State that the test was performed without independent verification by 
	Ziff; and 

G.      Follow proper trademark usage and acknowledge Ziff's trademark rights 
	(e.g., "[ ] achieved a score of X DOSMark (TM) units on the 
	PC Bench (TM) 8.0 test. PC Bench and DOSMark are trademarks of 
	Ziff-Davis Publishing Company, L.P."). 

This Agreement and your rights hereunder shall automatically terminate if you 
fail to comply with any provision of this Agreement. Upon such termination, 
you agree to cease all use of the Software, to delete the Software, and to 
destroy all copies of the diskette and other materials contained in this 
package in your possession or under your control, or, if downloaded, to 
destroy any and all copies of the Software in your possession or under your 
control.

2.      Additional Restrictions

A.      You shall not (and shall not permit other persons or entities to) 
	rent, lease, sell, sublicense, assign, or otherwise transfer the 
	Software or this Agreement. Any attempt to do so shall be void and of 
	no effect.

B.      You shall not (and shall not permit other persons or entities to) 
	reverse engineer, decompile, disassemble, merge, modify, include in 
	other software or translate the Software, or use the Software for any 
	commercial purposes, except for the publication of test results, as 
	provided above.

C.      You shall not (and shall not permit other persons or entities to) 
	remove or obscure Ziff's copyright, trademark or other proprietary 
	notices or legends from any of the materials contained in this 
	package or downloaded.

3.      Limited Warranty and Limited Liability

THE SOFTWARE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS 
OR IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTY OF MERCHANTABILITY 
AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE RESULTS AND 
PERFORMANCE OF THE SOFTWARE IS ASSUMED BY YOU, AND ZIFF ASSUMES NO 
RESPONSIBILITY FOR THE ACCURACY OR APPLICATION OF OR ERRORS OR OMISSIONS IN 
THE SOFTWARE. IN NO EVENT SHALL ZIFF BE LIABLE FOR ANY DIRECT, INDIRECT, 
SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR 
INABILITY TO USE THE SOFTWARE, EVEN IF ZIFF HAS BEEN ADVISED OF THE 
LIKELIHOOD OF SUCH DAMAGES OCCURRING. ZIFF SHALL NOT BE LIABLE FOR ANY LOSS,
DAMAGES OR COSTS, ARISING OUT OF, BUT NOT LIMITED TO, LOST PROFITS OR 
REVENUE, LOSS OF USE OF THE SOFTWARE, LOSS OF DATA OR EQUIPMENT, THE COSTS OF 
RECOVERING SOFTWARE, DATA OR EQUIPMENT, THE COST OF SUBSTITUTE SOFTWARE OR 
DATA, CLAIMS BY THIRD PARTIES, OR OTHER SIMILAR COSTS.

THE ONLY WARRANTY MADE BY ZIFF IS THAT THE ORIGINAL PHYSICAL MEDIA IN WHICH 
THE SOFTWARE IS EMBODIED AND WHICH IS DISTRIBUTED BY ZIFF SHALL BE FREE OF 
DEFECTS IN MATERIALS AND WORKMANSHIP. ZIFF'S ENTIRE LIABILITY AND THE USER'S 
EXCLUSIVE REMEDY SHALL BE LIMITED TO THE REPLACEMENT OF THE ORIGINAL PHYSICAL 
MEDIA IF DEFECTIVE. THE WARRANTIES AND REMEDIES SET FORTH HEREIN ARE 
EXCLUSIVE AND IN LIEU OF ALL OTHERS, ORAL OR WRITTEN, EXPRESS OR IMPLIED. NO 
ZIFF AGENT OR EMPLOYEE, OR THIRD PARTY, IS AUTHORIZED TO MAKE ANY 
MODIFICATION OR ADDITION TO THIS WARRANTY.

SOME STATES DO NOT ALLOW EXCLUSION OR LIMITATION OF IMPLIED WARRANTIES OR 
LIMITATION OF LIABILITY FOR INCIDENTAL OR CONSEQUENTIAL DAMAGES; SO THE ABOVE 
LIMITATIONS OR EXCLUSIONS MAY NOT APPLY TO YOU.

4.      U.S. Government Restricted Rights

The Software is licensed subject to RESTRICTED RIGHTS. Use, duplication or 
disclosure by the Government or any person or entity acting on its behalf is 
subject to restrictions as set forth in subdivision (c)(1)(ii) of the Rights 
in Technical Data and Computer Software Clause at DFARS (48 CFR 252.227-7013) 
for DoD contracts, in paragraphs (c)(1) and (2) of the Commercial Computer 
Software-Restricted Rights clause in the FAR (48 CFR 52.227-19) for civilian 
agencies, on in other comparable agency clauses. The contractor/manufacturer 
is the Ziff-Davis Benchmark Operation, One Copley Parkway, Suite 510, 
Morrisville, North Carolina 27560. 

5.      General Provisions

Nothing in this Agreement constitutes a waiver of Ziff's rights under U.S 
copyright laws or any other Federal, state, local or foreign law. You are 
responsible for installation, management, and operation of the Software. This 
Agreement shall be construed, interpreted and governed under New York law. If 
any provision of this Agreement shall be held by a court of competent 
jurisdiction to be illegal, invalid or unenforceable, the remaining 
provisions shall remain in full force and effect. 


Trademarks

Above DISK (R) is a registered trademark of Above Software.

Computer Select (R), Computer Shopper (R), PC/Computing (R), PC Direct (R), 
PC Expert (R), PC Magazine (R), PC Week (R), and WinBench (R) are registered 
trademarks and DOSMark (TM), First Looks (TM), MacBench (TM), NetBench (TM), 
PC Bench (TM), PC Direkt (TM), PC Professionell (TM), ServerBench (TM), 
Windows (TM) Activ (TM), and Winstone (TM) are trademarks of Ziff-Davis 
Publishing Company, L.P.

CompuServe (R) is a registered trademark of CompuServe, Inc.

dBASE (R) and Paradox (R) are registered trademarks of Borland 
International, Inc.

Harvard Graphics (R) is a registered trademark of Software Publishing Corp.

Intel (R) is a registered trademark of Intel Corporation.

Lotus (R) and 1-2-3 (R) are registered trademarks of Lotus Development 
Corporation.

Microsoft (R) and MS-DOS (R) are registered trademarks and Windows (TM) and 
Windows (TM) Sources (TM) are trademarks of Microsoft Corporation.

Micro Channel (TM) is a trademark of International Business Machines 
Corporation.

SCO (R) is a registered trademark of The Santa Cruz Operation, Inc.

UNIX (R) is a registered trademark of UNIX System Laboratories, Inc.

WordPerfect (R) is a registered trademark of WordPerfect Corporation.



Preface


This manual tells you how to install, set up, and run PC Bench (TM) on your 
PC and how to interpret your results.

In this preface, you can find:

+       A list of notes for this release of PC Bench.

+       A reading path telling you which chapters to read if you wish to 
	understand the concepts behind PC Bench and if you need to run 
	PC Bench and understand your results.

+       A summary of the chapters in this manual.

+       A list of conventions PC Bench and this manual use.


Introducing PC Bench

The Ziff-Davis Benchmark Operation (ZDBOp) developed PC Bench 8.0 as part of 
a family of benchmarks to help you evaluate your PC's performance. Each 
Ziff-Davis benchmark lets you test different aspects of a PC's performance.

PC Bench is a subsystem-level benchmark you can use to measure the 
performance of your PC's processor, memory, disk, and video subsystems. The 
scores you get with PC Bench give you a measure of the performance of these 
subsystems when you run DOS applications. PC Bench's DOSMark (TM) provides a 
general, overall measure of your PC's performance under DOS.

You can get copies of PC Bench and other Ziff-Davis benchmarks free of charge 
by contacting the Ziff-Davis Benchmark Operation (ZDBOp) or by downloading 
the benchmark you want from ZiffNet. To request benchmarks from ZDBOp, use 
the form included at the back of this manual. If you have technical questions 
about PC Bench, see Appendix C for information on how to contact ZDBOp.


Notes about PC Bench

Please note the following before you run PC Bench:

1.      You must read and agree to the license information in the front of 
	this manual before running PC Bench. The same licensing information 
	appears on your PC's screen the first time you run PC Bench after you 
	install it and in the README.TXT file included with PC Bench. If you 
	do not agree to the licensing information, delete all copies of 
	PC Bench. If ZDBOp sent you the PC Bench diskette, return it and all 
	accompanying materials (including any documentation) to ZDBOp at the 
	following address:

	Ziff-Davis Benchmark Operation
	One Copley Parkway, Suite 510
	Morrisville, NC 27560

	For more information on licensing PC Bench, see Chapter 2.

2.      PC Bench version 8.0 is the most recent release. This release differs 
	from previous releases in many important ways. As a result, you can 
	only meaningfully compare PC Bench 8.0 results with other PC Bench 
	version 8.0 results. Do not attempt to compare version 8.0 results 
	with results from earlier versions of PC Bench. Also, the database 
	format for PC Bench 8.0 is not compatible with the database formats 
	of earlier versions of PC Bench.

3.      PC Bench runs DOS-based tests. Its results are not meant to reflect 
	your PC's performance when running Windows (TM)-based applications.

4.      ZDBOp tested PC Bench 8.0 with the EMM386 and HIMEM.SYS memory 
	managers that ship with DOS. While PC Bench may run if you have 
	another memory manager loaded on your PC, ZDBOp does not recommend 
	using PC Bench as a tool for testing or comparing the performance 
	of memory managers.


Who should read this manual

This manual is for people who want to run PC Bench and understand what its 
scores mean about the overall performance of their PCs. If you use PC Bench, 
you should already be familiar with PCs in general. (If you have questions 
about your PC or the applications you run on it, consult the documentation 
that came with your PC or the applications.)

You can easily find the information about PC Bench you need in this manual. 
Early chapters explain PC Bench basics, and later chapters provide a more 
detailed explanation of how PC Bench works. This organization lets you choose 
the degree of technical information you need without having to read 
everything in the manual.

In addition, you can find the tasks you want to perform without reading the 
entire manual. For example, if you just want to view PC Bench results, read 
Chapter 4. The table that follows lists some general concepts and tasks you 
may need to understand and which chapter contains information about them.

If you need to: Read these chapters:

(Concepts)
Get a general overview of PC Bench              Chapter 1
Learn about the PC Bench tests                  Chapter 5
Understand what the scores mean                 Chapter 5
Understand ways to use your PC's scores         Chapter 5
Learn about what affects your PC's scores       Chapter 6
Understand the concepts behind PC Bench         Chapter 7
Understand where to find other PC Bench results Chapter 8

(Tasks)
Determine the requirements to run PC Bench      Chapter 2
Install and set up PC Bench                     Chapter 2
License PC Bench                                Chapter 2
Start PC Bench                                  Chapter 3
Run the PC Bench tests                          Chapter 3
Change your PC's test settings                  Chapter 3
Create a test suite                             Chapter 3
Save results                                    Chapter 4
Compare results                                 Chapter 4
Select a baseline system                        Chapter 4
Print results                                   Chapter 4
Use functions in the PC Bench program           Chapters 3, 4
View results from a test run                    Chapter 4
Use PC Bench with other Ziff-Davis benchmarks   Chapters 1, 5
Get tips on running PC Bench                    Chapter 6
Get a list of the PC Bench tests                Appendix A
Recover from a PC Bench error message           Appendix B
Get technical support from ZDBOp                Appendix C
Request Ziff-Davis benchmarks                   Appendix C,
						Benchmark Request Form


Summary of this manual

The following list contains a summary of the information in each chapter of 
this manual.

Chapter 1       An Overview of PC Bench
		Presents a brief description of PC Bench and gives an 
		overview of the scores PC Bench produces.

Chapter 2       Installing PC Bench
		Gives you the software and hardware requirements for 
		PC Bench. Explains how to install PC Bench and set up your 
		PC to run the benchmark. Also tells you how to license and 
		register your copy of PC Bench.

		NOTE:   In most cases, you must install PC Bench before you 
			can read or print this manual. If, however, you 
			received a copy of the manual without first 
			installing PC Bench on your PC, you can use the 
			instructions in this chapter to install it.

Chapter 3       Running PC Bench
		Tells you the different ways to run the PC Bench tests on 
		your PC. Explains how to run PC Bench in demo and batch 
		modes. Gives instructions on how to create your own custom 
		test suites.

Chapter 4       Working with PC Bench's Results
		Tells you how to use the PC Bench program screen to view 
		results and compare results. Explains how to save and print 
		results. Tells you how to import results from other PCs.

Chapter 5       Understanding PC Bench's Results
		Explains in detail the different scores PC Bench produces.

Chapter 6       What Affects PC Bench's Results
		Provides tips on configuring your PC for optimal performance 
		and tells you some factors that can influence your PC Bench 
		scores.

Chapter 7       The Concepts Behind PC Bench
		Explains how PC Bench measures the performance of your PC. 
		Discusses the types of tests PC Bench runs.

Chapter 8       How Your PC Compares
		Tells you where to find published results and standards to 
		use in comparing the PC Bench scores for your PC.

Appendix A      The PC Bench Tests
		Gives you a list of all the tests that come with PC Bench.

Appendix B      Troubleshooting
		Tells you what to do if you must abort PC Bench while it is 
		running a test. Lists some PC Bench error messages and tells 
		you how to deal with those errors.

Appendix C      Technical Support
		Explains how to get technical support for PC Bench.

Benchmark Request Form

Ziff-Davis Benchmarks' Manual Order Form


On-line documentation

PC Bench comes with the following on-line documentation:

README.TXT      This file tells you how to license PC Bench and contains 
		notes on PC Bench. You should read this information before 
		you run PC Bench.

PC Bench also includes two versions of this manual. They are:

PCBENCH.DOC     Microsoft (R) Word for Windows 2.0c version.

PCBENCH.TXT     ASCII text version; it does not contain any special 
		formatting commands. (This file contains special touch-up 
		work to help you print it.)

With these two versions, you should be able to print a copy of this manual 
regardless of the type of software you have on your PC. Note, however, that 
the PCBENCH.TXT file is not formatted and does not contain illustrations of 
sample screens.

NOTE:   If you would like to order a printed copy of this manual, see the 
	manual order form at the back of this book.

In addition to these files, PC Bench includes an on-line Help system that 
contains most of the information in this manual. To access the PC Bench help 
topics, select Help from the main menu in the PC Bench program screen.


Conventions PC Bench uses

When you work with the PC Bench program screen, you can use some standard 
conventions with which you may already be familiar. For example, the PC Bench 
main menu bar uses drop-down menus to present options. When you select a 
menu title, such as "File," a drop-down menu appears and displays the menu 
items available. In many cases, the drop-down menus duplicate actions you can 
perform using PC Bench's function buttons in the PC Bench program screen. In 
some cases, the drop-down menus give you additional options, such as getting 
information on how to contact ZDBOp.

PC Bench also uses several other standard conventions for selecting items 
from its main screen, windows, and dialog boxes. You can run PC Bench with or 
without a mouse attached to your PC. The following sections describe how to 
select items with a mouse and with keyboard commands.


Using a mouse

If you have a mouse attached to your PC, you can use it to select items by 
placing your cursor on the item you wish to select and then clicking your 
left mouse button. Generally, if you click on an item such as a function 
button or a test name in a list, you are selecting that item and PC Bench 
highlights it. If there is an action associated with the item you select 
(such as the Run function button), PC Bench takes the action associated with 
that item. For example, if you click on the PC Bench About function button, 
PC Bench displays a screen that tells you the names of the PC Bench 
development team.


Using keyboard commands

If you do not have a mouse attached to your PC or if you prefer not to use a 
mouse, you can:

+       Press the Alt key and the underlined access character of the item you 
	want to select. (For example, to select the Print option in the File 
	menu, press the Alt key, the "f" key, and then the "p" key.) This 
	item can be one of the PC Bench function buttons, a menu bar title, 
	or a menu item within a drop-down menu.

+       Press the Tab key or the Tab-Shift key combination to move the 
	current focus from function to function on the PC Bench program 
	screen. (You can identify the current focus by looking for the 
	highlighted or outlined item on the screen.) PC Bench highlights a 
	menu item by shading that item and highlights a function button by 
	outlining the button with a dotted line. The status line at the 
	bottom of the screen describes the current focus.

	When PC Bench highlights the item or function you want, press Enter 
	to select that item. For example, if the current focus is on Help in 
	the menu bar, when you press the Tab key, the focus goes to the About 
	function button. You would then press Enter to bring up the About 
	PC Bench screen. To de-select an item, press the Esc key.

+       Use the arrow keys to highlight a menu, submenu, or list item and 
	then press the Enter key to select that item. For example, if you 
	select the list of tests beside the Run function button on the 
	PC Bench program screen, you can then use your PC's arrow keys to 
	move the highlight in that list up or down. Once you highlight the 
	test you wish to run, press the Enter key to select that test.


General manual conventions

This manual uses a few general conventions. The manual:

+       Portrays information you enter from your keyboard and 
	system responses in a monospace or typewriter font. For example:

	C:\PCBENCH\BENCH.EXE

+       Uses italics to indicate variables where you must supply a value. 
	For example:

	C:\PCBENCH\BENCH.EXE \Ofile
	
	indicates that you should enter a DOS file name after the \O option.

+       Uses the phrase "your PC" when talking about the PC you are testing.

+       Uses the term "highlight" an item to mean "move the focus to" an 
	item. When you use keyboard commands to work with PC Bench, the item 
	you wish to choose must be active before you take an action on that 
	item. When an item is active, the focus for the screen is on that 
	item. PC Bench highlights items in its screen by either outlining the 
	item with a broken line (such as the Run function button) or shading 
	the item (as with the File menu title in the menu bar).

+       Uses all capital letters to denote DOS file names and directories. 
	For example, this manual contains sentences similar to the following:

	Before you install PC Bench, you should create a directory on your 
	hard drive called \PCBENCH to hold the PC Bench files.

+       Separates the keys in the notation of a key combination with a hyphen 
	(-). This means you do not have to press all the keys at the same 
	time. For example, to select Save from the File menu, you would use 
	the Alt-f-s key combination. You would first press the Alt key, then 
	the "f" key, and finally, the "s" key.

+       Uses the term "select" to tell you when to click on an item using 
	your PC's mouse or use a key combination to select an item. You can 
	use your PC's mouse or keyboard to select items as the previous 
	section describes.

+       Uses the following symbols to make it easier for you to know how to 
	perform functions with and without a mouse:

	{SYMBOL 56 \f "Wingdings"}      If you have a mouse attached to your 
		PC, look for this symbol before instructional paragraphs. 
		This symbol denotes paragraphs that tell you how to execute 
		PC Bench functions using your mouse. (ZDBOp recommends you 
		use PC Bench with a mouse if you have one available.)

	{SYMBOL 58 \f "Wingdings"}      If you do not have a mouse attached 
		to your PC, look for this symbol before instructional 
		paragraphs. This symbol denotes paragraphs that tell you 
		how to execute PC Bench functions from your keyboard.

	If an instructional paragraph does not have one of these symbols, the 
	paragraph will use the term "select" as in the bulleted item above. 
	In this case, you should use the method appropriate for your PC to 
	select the item.


Contacting ZDBOp

You can mail or fax your questions about PC Bench to ZDBOp, or you can use 
ZiffNet, the Ziff-Davis on-line service to contact ZDBOp. Appendix C contains 
details on the different ways to contact ZDBOp.


End of Preface


Table of Contents

Chapter 1:      An Overview of PC Bench       1
		The PC Bench basics     1
		How PC Bench measures a PC's performance        2
		Application profiling   3
		Subsystem scores and the DOSMark        3
		The subsystems PC Bench measures        4
		Understanding the tests 5
		The DOSMark     5
		The Processor Score     5
		The Video Score 6
		The Disk Score  6
		Using PC Bench results  7
		Understanding the baseline system       8
		Using PC Bench with other Ziff-Davis benchmarks 8
		What's next     9

Chapter 2:      Installing PC Bench 11
		Hardware and software requirements      11
		Installing PC Bench     12
		Setting up your PC to run PC Bench      13
		Licensing PC Bench      15
		What's next     16

Chapter 3:      Running PC Bench    17
		What you should know before you run PC Bench    17
		Changing your PC's test settings        18
		Starting PC Bench       22
		Using PC Bench command-line switches    23
		Exiting PC Bench        25
		Using the PC Bench program screen       26
		Selecting tests 28
		Running the PC Bench tests      30
		Running the DOSMark Test Suite  32
		Running other PC Bench built-in test suites     32
		Running all the PC Bench tests  33
		Running selected PC Bench tests 33
		Running PC Bench in batch mode  34
		Running PC Bench in demonstration mode  36
		Running the tests included as utilities 37
		Saving your test results        38
		Creating and running custom test suites 38
		Creating a custom test suite    38
		Editing a custom test suite     40
		Using previously created test suites    41
		What's next     42

Chapter 4:      Working with PC Bench's Results     43
		Understanding the graph buttons 44
		Comparing results       47
		Selecting a baseline system     48
		Adding a set of results to the display  50
		Importing PC Bench results      51
		Removing a set of results from the displayed results    54
		Deleting results from the PC Bench database     56
		Viewing your PC's configuration information     57
		Printing results        59
		What's next     61

Chapter 5       Understanding PC Bench's Results       63
		The PC Bench scores     63
		The DOSMark     64
		The Processor Score     65
		The Video Score 65
		The Disk Score  66
		The Memory Tests results        66
		Other PC Bench tests    67
		Why results differ      68
		Publishing PC Bench results     69
		Using PC Bench results  69
		Using PC Bench with other benchmarks    70
		What's next     71

Chapter 6:      What Affects PC Bench's Results     73
		Software and hardware factors that influence PC Bench's results 73
		Tips for improving your PC Bench results        75
		Following proper testing procedures     75
		Other ways to improve your PC Bench results     76
		Using PC Bench to best advantage        76
		What's next     77

Chapter 7:      The Concepts Behind PC Bench        79
		The design goal 79
		Profiling applications  80
		Understanding the "market-centered" concept     81
		Using a weighted harmonic mean  81
		Understanding the Disk Mix      82
		PC Bench's Protected Mode Large Instruction Mix 82
		Correlating the score to applications   83
		What's next     83

Chapter 8:      How Your PC Compares        85
		Checking published results      85
		Using PCs within your organization      86
		Publishing PC Bench scores      87
		What's next     87

Appendix A:     The PC Bench Tests 89
		Processor tests 89
		EGA/VGA text mode tests 91
		EGA/VGA graphics mode tests     92
		Disk tests      93
		Disk throughput tests   93
		DOS disk access 99
		Memory timing tests     99
		Tests included as utilities     102
		Battery Rundown Test    102
		Time the Timer Test     102
		VGA Compatibility Test  102
		EGA/VGA Monitor Quality Test    103

Appendix B:     Troubleshooting    105
		What to do if you have problems 105
		Running PC Bench with memory managers   107
		Running PC Bench in batch mode  107
		PC Bench error messages 108
		General database errors 112

Appendix C:     Technical Support  113
		Requesting a benchmark  113
		Ordering a printed manual       114
		Contacting ZDBOp        114

Benchmark Request Form

Ziff-Davis Benchmarks' Manual Order Form

Acknowledgments




Chapter 1
An Overview of PC Bench



This chapter presents a brief description of PC Bench and the results it 
produces. After reading this chapter, you should have a basic understanding 
of how PC Bench measures your PC's performance and what its results mean.

NOTE:   PC Bench is available free of charge on a 3.5" diskette. To request 
	a copy of PC Bench or any other Ziff-Davis benchmark, use the form at 
	the back of this manual. You can also download the benchmark you want 
	from ZiffNet. See Appendix C for more information.


The PC Bench basics

PC Bench is a subsystem-level, synthetic benchmark you can use to measure the 
performance of your PC's processor, memory, disk, and video subsystems. 
PC Bench's results give you a measure of the performance you can expect from 
your PC when you run DOS applications. PC Bench's DOSMark provides an overall 
measure of your PC's performance. The PC Bench tests reflect the most common 
operations that best-selling DOS applications perform.

ZDBOp developed the PC Bench tests by profiling such applications to 
determine the types of operations the applications typically perform. Thus, 
PC Bench's results correlate closely with the performance of actual 
applications.

PC Bench's user interface lets you enter information about your PC, run 
tests, create test suites, and compare results. You perform most PC Bench 
functions from the PC Bench program screen. Once you run a test, PC Bench 
displays its results in the PC Bench program screen in graph format. (For 
more information on how to use the PC Bench program screen, see Chapters 3 
and 4.)

NOTE:   PC Bench version 8.0 is the most recent release of this product. This 
	release differs from previous releases in many important ways. As a 
	result, you can only meaningfully compare PC Bench  version 8.0 
	results with other PC Bench version 8.0 results. Do not attempt to 
	compare your PC's version 8.0 results with results from earlier 
	versions of PC Bench. Also, the database format for PC Bench 8.0 is 
	not compatible with the database formats of earlier versions of 
	PC Bench.


How PC Bench measures a PC's performance

PC Bench tests your PC's processor and disk subsystems using test mixes of 
operations that closely reflect the typical operations DOS applications 
perform. A "mix" is a single test consisting of a combination of operations. 
PC Bench's mixes use operations that subsystems execute when running 
applications.

NOTE:   The PC Bench video subsystem tests do not use mixes; they report raw 
	performance results for different VGA video operations. PC Bench then 
	weights the most important video results and includes them in its 
	DOSMark.

ZDBOp modeled the PC Bench tests on operations like those generated by DOS 
applications. PC Bench does not run applications during its tests.

The sections below explain briefly how the PC Bench tests measure your PC's 
performance. For more information on the types of tests in PC Bench, see the 
section "Understanding the tests" later in this chapter.


Application profiling

The individual tests that PC Bench contains reflect extensive profiling of 
leading DOS applications. Profiling is a method of monitoring an application 
to see the types of operations it performs. By using application profiles as 
targets, PC Bench's tests can accurately reflect what the applications do. 
Extensive profiling research provided information on processor, memory, disk, 
and video access patterns in complex usage scenarios in actual applications. 
ZDBOp used the information from profiling to improve PC Bench's testing 
techniques with such features as:

+       A new 16-bit Protected Mode Large Instruction mix. As applications 
	have changed, PC Bench's instruction-mix technology has changed with 
	them. The large working set size (approximately 450 KB) of this mix 
	lets it accurately reflect application CPU and RAM usage.

+       A new Disk Mix technology. PC Bench's Disk Mix accurately reflects 
	the way today's applications use a PC's disk subsystem. The Disk Mix 
	performs a broad range of disk operations over a large set of files. 
	The Disk Mix works well and accurately with the software and hardware 
	disk caches used in today's PCs, including such disk caching tools as 
	SMARTDRV.

+       More accurate weights that provide realistic measures of a PC's 
	performance. For example, the individual video tests weight 8- and 
	16-bit writes to both the text and graphics region of VGA display 
	memory. These tests provide an accurate portrayal of the way DOS 
	applications actually use a PC's video subsystem.


Subsystem scores and the DOSMark

PC Bench provides four main scores you can use to measure your PC's 
performance:  the DOSMark score and the Processor, Video, and Disk subsystem 
scores. PC Bench produces each subsystem score from a different set of tests. 
(For more information on these tests, see the next section, "Understanding 
the tests.")

The DOSMark is a weighted, harmonic mean of the Processor, Video, and Disk 
subsystem scores. PC Bench also weights the results of the scores of its 
individual tests to produce the subsystem scores. Using a harmonic mean 
provides a way to aggregate the scores for the individual tests into a 
weighted, overall score. PC Bench measures the total number of test 
operations and then divides that number by the total time it takes to perform 
these operations. PC Bench then weights this number to account for the 
different levels of importance each operation within a test group has. 
Basically, the more frequently an application uses an operation, the more 
weight PC Bench gives its test for that operation.

NOTE:   PC Bench also reports an absolute value for each individual test you 
	run.


The subsystems PC Bench measures

PC Bench measures your PC's processor, memory, disk, and video subsystems. 
Each of these subsystems includes different components. In evaluating these 
subsystems, PC Bench considers both the software and hardware that make up 
the subsystem.

The processor and memory subsystems include the main processor chip 
(for example, Intel{SYMBOL 226 \f "Symbol"} 80486 DX), any internal or 
external cache and cache controller, and main memory. The processor subsystem 
also includes the math co-processor if your PC has one.

The disk subsystem includes the hard disk, the disk controller, and any 
hardware and software disk caches your PC uses.

Finally, the video subsystem is the video adapter, whether on the motherboard 
or a plug-in card.


Understanding the tests

The following sections explain briefly the different PC Bench built-in test 
suites and the scores these tests produce. (A test suite is a group of tests 
you run at once with a single command.) For complete information on the 
different PC Bench tests you can run and what the different scores mean about 
your PC's performance, read Chapter 5. (If you need to run the PC Bench 
tests, skip to Chapter 3.)


The DOSMark

When you execute the DOSMark Test Suite, PC Bench runs the tests necessary to 
produce the Processor, Video, and Disk scores and then calculates the 
weighted harmonic mean of these three scores to produce a single number 
called the DOSMark. The DOSMark provides an overall measure of your PC's 
performance when you run DOS applications on your PC. You can use the DOSMark 
to compare your PC with other PCs.


The Processor Score

The core of PC Bench's Processor Score is the 16-bit Protected Mode Large 
Instruction Mix. This new Large Instruction Mix reflects the larger 
working-set sizes of today's applications. The approximate working-set size 
of this mix is 450 KB. ZDBOp designed this mix to accurately simulate the way 
popular DOS applications use a PC's processor. The 16-bit Protected Mode 
Large Instruction Mix reflects the way applications stress a PC's CPU 
subsystem in such areas as bus utilization, CPU cache usage, CPU cache 
interface, external cache usage, main memory, and the CPU instruction set.

PC Bench's built-in Processor Test Suite also includes tests for the floating 
point unit (FPU).

When you run this suite, PC Bench takes the results of the CPU and FPU tests 
and calculates a weighted harmonic mean. The weighted result reflects the 
relative effect of a PC's CPU and FPU on typical applications.

NOTE:   PC Bench includes a math co-processor test that requires an FPU. If 
	your PC does not have an FPU, PC Bench uses FPU-emulation for this 
	particular test. Results with FPU-emulation will be significantly 
	slower than those with an FPU present.


The Video Score

The Video Score is a combination of different individual video tests. The 
Video Test Suite uses low-level video tests to accurately measure how your 
PC's video subsystem performs when running applications under DOS. When you 
run the Video Test Suite, PC Bench takes the results for the individual video 
tests in the test suite and calculates the weighted harmonic mean of those 
results. The resulting score is your PC's Video Score. The individual video 
tests measure 8- and 16-bit VGA write operations involving both text and 
graphics. The Video Score is a single number that provides an overall measure 
of your PC's DOS video performance.


The Disk Score

The Disk Score lets you gauge the performance of your PC's disk subsystem. 
That subsystem includes the hard disk, disk controller, and any hardware or 
software disk caches. PC Bench 8.0 uses a new disk-testing methodology:  its 
Disk Mix.

The PC Bench Disk Mix performs a range of disk operations over a large set of 
files. This mix uses over 100 files spread across six directories. The Disk 
Mix works with multiple files concurrently, creates and deletes files, moves 
data in groups of different sizes, intersperses write and read operations, 
and generally uses the disk subsystem the way applications do. The Disk Mix 
works with a PC's disk subsystem via the standard Int 21H interface, so you 
can use the Disk score as an accurate measure regardless of whether you are 
using a software or hardware disk cache.

ZDBOp based PC Bench's Disk Mix on disk usage in profiled applications. Thus, 
the Disk Mix works with your PC's disk subsystem the way applications do.


Using PC Bench results

PC Bench produces a result for each individual test it runs even if that test 
is part of a test suite. For example, if you run the DOSMark Test Suite, you 
will get an absolute result for each individual test in the suite in addition 
to the DOSMark for your PC. You can use the different results PC Bench 
produces to compare your PC with others and to see how well your PC handles 
certain operations.

PC Bench stores its results in a database, and you can use results in this 
database for comparison. This ability lets you:

+       Run PC Bench multiple times on the same PC with different 
	configurations and compare the results for each test run. For 
	example, you could run PC Bench on a PC with and without software 
	disk caching and compare the disk scores.

+       Compare PC Bench results for your PC to the results for other PCs. 
	One way to compare PC Bench results for different PCs is to use 
	published results from Ziff-Davis publications. (For a list of these 
	publications, see Chapter 8.) Also, your organization may have more 
	than one type of PC or may have PCs with different configurations. If 
	so, you can run PC Bench on the different machines within your 
	organization to compare their performance. For more information on 
	using results from different PCs, see the section "Importing PC Bench 
	results" in Chapter 4.

NOTE:   If you want to compare different PCs in the same way, make sure all 
	the PCs you want to compare were set up the same way. For example, 
	make sure you do not compare the results of a PC that had disk 
	caching enabled and a PC that did not unless you are trying to gauge 
	the effect of the disk cache. Otherwise, your comparison may not 
	reflect the true potential of both systems fairly.

For more suggestions on ways to use your PC Bench results, see the section 
"Using PC Bench results" in Chapter 5.


Understanding the baseline system

When you display results in the PC Bench program screen, you can choose 
which set of results you want to use as a baseline system. Once you set a 
baseline system, this is your comparison system for all other sets of 
results. PC Bench sets the baseline system's results to 100% and displays all 
other results as a percentage of the baseline system's results. When you use 
a baseline system, you can quickly determine from the displayed graphs how 
results for different PCs compare. For complete information on how to set a 
baseline system, see the section "Selecting a baseline system" in Chapter 4.


Using PC Bench with other Ziff-Davis benchmarks

Although PC Bench tests the speed of a PC's processor, memory, disk, and 
video subsystems to determine how they perform when they use DOS 
applications, it does not give a complete measure of your PC's performance if 
you also use Windows-based applications. To get that information, you can use 
PC Bench in combination with two other Ziff-Davis benchmarks: WinBench (R) 
and Winstone (TM) 94.

+       You can use WinBench to measure the performance of your PC's graphics 
	and disk subsystems under Windows. ZDBOp based the WinBench tests on 
	profiles of leading Windows applications. Thus, the WinBench tests 
	provide an accurate measurement of your PC's graphics and disk 
	performance under Windows.

+       Use Winstone 94 to measure the overall performance of your PC as it 
	runs actual Windows-based applications. With Winstone 94 you can 
	measure the effect that changing individual system components has on 
	your PC's overall performance when running Windows and Windows-based 
	applications. Winstone 94 performs tasks within the applications it 
	runs like those typical users perform with those applications. 
	Winstone 94 is the sole Microsoft-certified benchmark for the testing 
	of industry-standard PCs running Windows 3.1-based applications.

You can use the form at the back of this manual to request copies of 
WinBench, Winstone 94, and other Ziff-Davis benchmarks from ZDBOp. You can 
also download Ziff-Davis benchmarks from ZiffNet. For more information, see 
Appendix C.


What's next

Now that you have an overview of PC Bench and how it tests a PC, read 
Chapter 2 to install PC Bench and set up your PC to run the benchmark.

Chapters 3 and 4 give details on how to run the PC Bench tests and how to 
view and compare results.

If you would like further details on the PC Bench tests and what the results 
mean about your PC's performance, read Chapter 5.


End of Chapter



Chapter 2
Installing PC Bench


This chapter tells the hardware and software requirements your PC must meet 
to run PC Bench successfully. It also explains how to set up your PC to run 
the benchmark and tells you how to install and license PC Bench. After you 
read this chapter and install PC Bench, you should be ready to run the 
PC Bench tests.

NOTE:   You can also learn how to install PC Bench by reading the 
	README.TXT file that comes with PC Bench. In most cases, you will 
	have installed PC Bench before printing a copy of this manual. 
	However, if you are installing PC Bench on another PC or if you 
	received a copy of this manual before you installed PC Bench, then 
	you can follow the installation instructions in this chapter.


Hardware and software requirements

PC Bench version 8.0 requires the following minimum hardware and software 
components:

+       A PC running MS-DOS{SYMBOL 226 \f "Symbol"} version 3.1 (or 
	compatible) or higher. For best results, use MS-DOS version 5.0 (or 
	compatible) or higher.

+       A 386 (or compatible) CPU or better.

	NOTE:   While you can run PC Bench on PC with a 286 CPU, you will not 
		be able to run all the PC Bench tests. PC Bench 8.0 requires 
		a minimum of a 386 CPU.

+       A minimum of 6.6 MB of free disk space to install the PC Bench files.

+       A minimum of 35 MB of free disk space on each drive you want to test.

+       A minimum of 543 KB of free conventional memory if you plan to run 
	all the PC Bench tests.

+       A minimum of 2 MB of extended memory to run the DOSMark tests.

+       A VGA video adapter.


Installing PC Bench

You can request a copy of PC Bench on a diskette from ZDBOp, or you can 
download it from ZiffNet, Ziff-Davis's on-line service. (See Appendix C for 
more information.)

PC Bench comes as a self-extracting file. To install PC Bench:

1.      Create a directory called \PCBENCH on your PC's disk.

	NOTE:   You do not have to create a separate PC Bench directory 
		called \PCBENCH, but ZDBOp recommends you do so.

2.      Change directories (CD) to the directory you just created.

3.      Put the PC Bench diskette in your PC's floppy drive and execute the 
	following command.

	PCBENCH.EXE

	For example, if your PC's floppy drive is the A drive, you would 
	type:

	A:\PCBENCH.EXE

	(If you downloaded PC Bench from ZiffNet, change directories to the 
	directory where you stored the PCBENCH.EXE file and then execute the 
	file.)

	PC Bench installs its files in the current directory.

4.      Type the following command from the directory where you 
	installed PC Bench to start PC Bench's database with a set 
	of sample comparison results:

	CLEARDB
	
	NOTE:   You do not have to execute the CLEARDB command to 
		run PC Bench. However, if you do not execute this 
		command, the PC Bench database will not contain any 
		sample comparison results when you start PC Bench. 
		Whether the database contains sample results will 
		not affect how PC Bench runs, but ZDBOp recommends 
		you execute this command. 



Setting up your PC to run PC Bench

Before you run PC Bench, you need to modify your CONFIG.SYS and AUTOEXEC.BAT 
files. To do this:

1.      Make backup copies of your original CONFIG.SYS and AUTOEXEC.BAT files 
	(you can use file names such as CONFIG.SAV and AUTOEXEC.SAV).

2.      Modify your CONFIG.SYS file so it contains the statement FILES=20. Do 
	not omit this statement. You can include a statement for a mouse 
	driver or any other drivers necessary for your PC to operate. You 
	should not load any non-essential Terminate and Stay Resident 
	programs (TSRs) or memory managers.

	NOTE:   While PC Bench may run if you have a memory manager loaded on 
		your PC, ZDBOp does not recommend using PC Bench as a tool 
		for testing or comparing the performance of memory managers. 
		If you run PC Bench with a memory manager loaded, and you 
		have problems running the tests, you should re-run the tests 
		without the memory manager loaded.

	The following is a sample minimal CONFIG.SYS file:

	FILES=20

	If you usually run DOS applications on your PC with HIMEM.SYS and DOS 
	loaded high (DOS=HIGH), then you should run PC Bench the same way. 
	For example, your CONFIG.SYS file would then be similar to the 
	following:

	DEVICE=C:\WINDOWS\HIMEM.SYS
	FILES=20
	DOS=HIGH

3.      Modify your AUTOEXEC.BAT file so it contains only the essential 
	drivers and programs necessary to operate your PC. For example:

	PATH=C:\DOS
	PROMPT $P$G
	C:\MOUSE\MOUSE

	You may also load a disk caching tool, such as SMARTDRV, from your 
	AUTOEXEC.BAT file.

	NOTE:   If you normally run DOS applications on your PC with a disk 
		caching tool enabled, you should run PC Bench the same way.) 
		Your file would then also contain a line similar to the 
		following:

		C:\WINDOWS\SMARTDRV.EXE

Remember, after you modify these system files you will need to reboot your PC 
for the changes you made to take effect. You can reboot your PC by pressing 
Ctrl-Alt-Del, by pressing the PC's reset button (if there is one), or by 
turning off the power to the PC and restarting it.


Licensing PC Bench

Before you run PC Bench, you must license and register it. By licensing 
PC Bench, you agree to the requirements and terms of its license agreement.

NOTE:   The first time you run PC Bench, it displays the PC Bench License 
	Agreement. (You will find the same license agreement at the front of 
	this manual and in the on-line README.TXT file.)

If you wish to display the license agreement at any time, go to the directory 
where you installed PC Bench and enter the command:

BENCH \L

PC Bench displays its license agreement on your PC's screen.

Figure 2-1:  The PC Bench License Agreement
{EMBED MSDraw   \* mergeformat|}

Use the space bar to page through the license agreement as you read it. Once 
you have read the entire agreement, PC Bench prompts you to register your 
copy of PC Bench by entering your name and your organization's name (if any). 
After you enter this information, press F10 or the Enter key to continue.

NOTE:   You only need to license your copy of PC Bench once.

If you choose not to adhere to the PC Bench License Agreement, delete all 
copies of PC Bench from your PC. If you loaded PC Bench from a diskette 
rather than by downloading it from ZiffNet, return the diskette and any 
documentation to ZDBOp at the following address:

Ziff-Davis Benchmark Operation
One Copley Parkway, Suite 510
Morrisville, NC 27560


What's next

Now that you have installed PC Bench, you can read Chapter 3 to learn how to 
run PC Bench's tests and how to change PC Bench's settings. Chapter 4 tells 
you how to view and compare PC Bench results.

Chapter 5 discusses the different PC Bench tests and what the scores mean 
about your PC.


End of Chapter



Chapter 3
Running PC Bench


This chapter explains what you need to know before you run PC Bench, how to 
change your PC's test settings, how to start PC Bench, and how to use 
functions in the PC Bench program screen. It also tells you how to run the 
PC Bench tests on your PC. After reading this chapter, you should know how to 
run PC Bench on your PC.


What you should know before you run PC Bench

Before you run PC Bench, make sure you:

+       Change your PC's CONFIG.SYS and AUTOEXEC.BAT files as described in 
	the section "Setting up your PC to run PC Bench" in Chapter 2.

+       License and register your copy of PC Bench.

Also, please note the following before you run any PC Bench tests:

1.      ZDBOp recommends you change the Machine ID and Machine Name fields in 
	PC Bench's Machine Information dialog box to names that are 
	meaningful to you. (For more information on how to change this 
	information, see the next section, "Changing your PC's test 
	settings.")

2.      PC Bench version 8.0 does not support EGA resolutions or EGA-only 
	adapters.

3.      If you wish to run all of PC Bench's tests in succession, note the 
	following:

	+       All disk tests will attempt to run on the same drive, because 
		PC Bench lets you set just one drive for the disk tests.

	+       To run either the Disk Throughput Tests or the Disk Mix 
		requires a minimum of 32 MB of free disk space.

	+       You will only get results for the BIOS Disk Access Tests if 
		the drive parameter in the Disk Settings section of the 
		Machine Information dialog box is set to a primary disk 
		partition (usually drive C:).

	+       The DOS File Access test will not run on a networked drive.

4.      If you have disk caching tools installed and enabled on your PC, your 
	PC Bench results for the disk tests will typically be better than 
	results without the disk caching. If you normally use your PC with a 
	disk caching tool enabled, then use the same tool with PC Bench and 
	it will give you a realistic measurement of how DOS applications 
	interact with your PC's disk. To get a true measurement of the 
	performance of just your hard disk itself, run PC Bench without disk 
	caching enabled.


Changing your PC's test settings

PC Bench lets you define information about your PC before you start a test 
run. You can set this information from within PC Bench using the Settings 
function button. By changing this information from test run to test run, you 
can easily identify different test runs.

To change your PC's test settings for a PC Bench test run, select the 
Settings function button.

{SYMBOL 56 \f "Wingdings"}      Click on the Settings function button in the 
	PC Bench program screen.

{SYMBOL 58 \f "Wingdings"}      Highlight the Settings function button using 
	the Tab key and then press the Enter key to select the button. (The 
	status message will say "Configure Machine ID and Disk Test 
	settings.") You can also use either the keyboard combination Alt-e to 
	select the Settings function button or the keyboard combination 
	Alt-s-t to select the Settings option from the Test Suites menu.

PC Bench displays the Machine Information dialog box.

Figure 3-1:  The Machine Information dialog box
{EMBED MSDraw   \* mergeformat|}

From the Machine Information dialog box you can define some of the settings 
PC Bench uses when it runs a test. PC Bench initially defines some of the 
settings in this box with default values. You can change any of the default 
values you wish. For example, you can set a machine name to make that 
machine's results easy to identify when you view and compare test results. 
(ZDBOp recommends you change at least the Machine ID and Variant fields for 
each test run.) Once you have the variables in this dialog box the way you 
want, select the Use button.

NOTE:   PC Bench uses the values you specify in the Machine Information 
	dialog box when you run tests. PC Bench does not, however, commit 
	these values to its database until you save results. If you change 
	your PC's settings before you save results, PC Bench updates all 
	unsaved results. Thus, when you do save the results, PC Bench commits 
	all the results to the database with the same settings. 
	
	Note also that after you save results or restart PC Bench, PC Bench 
	preserves all the settings from your last PC Bench session in the 
	Machine Information dialog box except the Variant fields. PC Bench 
	automatically changes the Variant fields to distinguish the current 
	results from previous test runs.

You do not have to enter any information in this dialog box to run PC Bench. 
Doing so, however, helps you distinguish between different PCs in the 
database and different set of results for a single PC. If you want to test 
the same PC under different configurations, you can use the Variant fields 
to identify each set of results. PC Bench uses the current time and date to 
initialize these fields for each new set of results. You can manually set one 
or more of the Variant fields using your own naming conventions. PC Bench 
then records the results as separate entries in its database.

The Machine Information dialog box contains two sections: Machine Settings 
and Disk Settings. The following list summarizes the fields in the Machine 
Settings section.

+       Machine ID is an alphanumeric entry up to 12 characters long you can 
	use to identify your PC. By default, this field is set to a string 
	that identifies your PC's CPU type (for example, 80486).

+       The five Variant boxes each contain an alphanumeric entry up to three 
	characters long. You can change these entries to make different 
	PC Bench test runs on the same PC easily identifiable. By default, 
	the values in these fields contain the date and time you either 
	started this session of PC Bench or last saved PC Bench results. 
	(The date and time format is mm dd hh mm ss.) For example, for a 
	date and time stamp of November 28, 15:30:10, the Variant fields 
	would be as follows:

	Variant 1 = 11
	Variant 2 = 28
	Variant 3 = 15
	Variant 4 = 30
	Variant 5 = 10

	You can change any of these fields to any entry up to three 
	characters long.

+       Machine Name is an alphanumeric entry up to 12 characters long you 
	can use to record the brand name and model of your PC. By default, 
	the entry in this field identifies your PC's CPU type (for example, 
	80486).

+       Project is a numeric entry up to seven characters long. If the tests 
	you are running are part of a particular project for your 
	organization, you may want to enter that project's name here, or you 
	can just use this field for additional descriptive information about 
	your PC.

+       Test Org is an alphanumeric entry up to 25 characters long containing 
	the name of the organization performing the test. This field defaults 
	to the organization name you entered when you licensed PC Bench.

+       Tester is an alphanumeric entry up to 25 characters long containing 
	the name of the person performing the test. This field defaults to 
	the licensee name you entered when you licensed PC Bench.

+       Machine Description is an alphanumeric string up to 224 characters 
	that can contain any pertinent information about your PC. To enter 
	information here, position your cursor in the text box and begin 
	typing.

The Machine Information dialog box also contains a section where you can 
specify settings for PC Bench's disk tests. In the Disk Settings section you 
can set the following fields:

+       Disk and Path. These are the disk and directory names where PC Bench 
	will build its test files for the Disk Mix tests. The disk where 
	PC Bench builds its test files is also the disk PC Bench tests. You 
	can select the disk and directory name using drop-down menus.

	{SYMBOL 56 \f "Wingdings"}      Click on the down-arrow beside a test 
		box to display the list of disks and directory names 
		available.

	{SYMBOL 58 \f "Wingdings"}      Highlight either the disk or path 
		text boxes using the Tab key and then press the down-arrow 
		key to display the list of disks and directory names 
		available. Use the up and down arrow keys to highlight your 
		selection and then press the Enter key to select the 
		highlighted item.

+       Repetitions. This tells PC Bench how many times to perform the disk 
	test(s). By default, this field is set to one. In almost all cases, 
	you will not need to increase the number of repetitions of the disk 
	tests; however, if you want PC Bench to repeat the disk tests, you 
	can type the number of repetitions you want in this text box.


Starting PC Bench

To run PC Bench, go to the directory where you installed PC Bench (ZDBOp 
recommends \PCBENCH) and enter the following command from the DOS command 
line:

BENCH

NOTE:   Before you run PC Bench, you must license it. The PC Bench License 
	Agreement and User Information screens appear the first time you 
	start PC Bench after installing it. Once you license and register 
	your copy of PC Bench, you do not need to do so again regardless of 
	how many times you start PC Bench. If you would like to re-license 
	your copy of PC Bench, you can start PC Bench with the \L option to 
	display the license agreement. For information on how to license and 
	register your copy of PC Bench, see the section "Licensing PC Bench" 
	in Chapter 2.

PC Bench briefly displays the Ziff-Davis benchmark stopwatch logo and 
PC Bench copyright information. PC Bench then displays its program screen.

NOTE:   If you would like to pass by these screens quickly, you can press the 
	Enter key once your PC begins to display the stopwatch logo or the 
	copyright information and PC Bench will move to the next screen.


Using PC Bench command-line switches

There are several command-line switches you can use with PC Bench. These 
switches let you direct PC Bench's actions. For example, you can use the /B 
option to run PC Bench in batch mode.

The PC Bench command-line switches are:

/? or /H        Displays a list of the switches PC Bench accepts and names 
		the DOS environment variables it uses.

/L              Displays the license agreement and lets you re-register 
		PC Bench. (PC Bench displays its license agreement 
		automatically the first time you run the benchmark if you 
		have not previously licensed and registered your copy of 
		PC Bench.)

/B              Runs all the PC Bench tests in batch mode using the current 
		DOS environment variables. You can set these variables from 
		the DOS command line when you start PC Bench or from a DOS 
		batch file.

		The variables you set for PC Bench when you run it in batch mode let you distinguish different sets of results. When you run PC Bench in batch mode, you can use the following DOS environment variables:
		
		+       PIN (machine ID)
		+       VARIANT1 through VARIANT5 (test variants)
		+       TYPE (machine name)
		+       PROJECT (project name)
		+       TESTORG (test organization)
		+       TESTER (tester's name)

		If you run PC Bench in batch mode multiple times on the same 
		PC, you can use these variables to identify different test 
		runs by changing variables from one test run to another. You 
		must set the PIN and VARIANTS DOS environment variables, and 
		these variables must be unique for each set of results.

/B /Ffile       Executes PC Bench in batch mode using the tests listed in the 
		test suite file. The file name must be a legal DOS file name 
		and must have a .STE extension. If the file is not in the 
		PC Bench directory, you must supply the full path name for 
		the file. You can create the test suite file from PC Bench's 
		program screen by using the Edit and Save options in the Test 
		Suites menu.

		NOTE:   If you use the /B option alone, PC Bench will run all 
			its tests. If you use the /B option with the /Ffile 
			option, PC Bench will run only the tests specified in 
			the test suite file.

/M              Prevents PC Bench from logging your PC's configuration 
		information. You should not stop PC Bench from logging your 
		PC's configuration information unless you are having problems 
		running PC Bench.

/Ofile          Logs your PC's results to a CSV (comma separated value) 
		format file called file. You can then use this file to view 
		or print test results. You can supply the path name where you 
		want PC Bench to save the file. The file name must be a legal 
		DOS file name with a .CSV extension.

/P              Prevents PC Bench from automatically determining what type of 
		processor is in your PC. This automatic determination can 
		cause problems on future PCs with new processors or on DOS 
		emulators.

NOTE:   These PC Bench command-line switches are not case sensitive.


Exiting PC Bench

You can exit the PC Bench program screen by:

+       Selecting the Exit function button.

+       Selecting Exit from the File menu.

+       Using the key sequence Alt-x.

NOTE:   If you have unsaved test results when you exit PC Bench, it will ask 
	if you want to save your results. You can select Save, Delete, or 
	Cancel in this warning box. If you choose to save your results, it 
	may take a few moments before PC Bench completes the save and the DOS 
	prompt appears. (For example, some laptops may take over one minute 
	to save test results and display the DOS prompt.)


Using the PC Bench program screen

When you start PC Bench, its program screen appears on your PC's screen.

Figure 3-2:  The PC Bench program screen
{EMBED MSDraw   \* mergeformat|}

This screen has three main parts: the menu bar at the top, the function 
buttons on the left side, and the results graph buttons on the right side. 
When you execute a test suite, a set of tests, or an individual test, 
PC Bench displays status messages in a fourth small area, the bar at the 
bottom of the screen, to let you know what it is doing.

You can select the actions you want to perform using either the drop-down 
menus from the menu bar or the function buttons.

NOTE:   When you work with the PC Bench program screen, you can use either 
	your mouse or your keyboard to select the function buttons and menu 
	bar items in the program screen. (It is usually easier to use 
	PC Bench with a mouse.)

	Note also that if you have not yet run any PC Bench tests or if you 
	have not run enough tests to compute the DOSMark or subsystem scores, 
	the graph buttons for the DOSMark and subsystem scores will read 
	"(Insufficient Data to Compute Score)". In addition, the Memory Tests 
	graph button will read "(No Results)" when there are no Memory Tests 
	results.

As PC Bench completes a test, it displays a horizontal bar graph of the 
result of that test. Once a test run is complete, PC Bench updates the 
DOSMark and subsystem scores and displays those scores on the results graph 
buttons in the main program screen. (For more information on how to work with 
and understand your PC's results, see Chapter 4.)

The PC Bench program screen includes the following function buttons:

+       About PC Bench. Selecting this button gives you information about the 
	PC Bench development team. You can also view this information by 
	selecting About PC Bench from the Help menu.

+       System Info. Selecting this button opens the System Information 
	dialog box that displays your PC's configuration. You can also 
	display the System Information dialog box by selecting Display System 
	Info from the File menu.

+       Run. Selecting this button starts the test suite visible in the 
	drop-down list beside the function button. You can display a list of 
	the available test suites by selecting the down arrow to the right of 
	this text box. Choose the test suite you want and then select the Run 
	button to execute it.

	NOTE:   If you want to run a single test or set of tests without 
		creating a test suite, go to the menu bar and select Choose 
		Test from the Test menu. For more information, see the 
		section "Running selected PC Bench tests" later in this 
		chapter.

+       Compare Results. Selecting this button displays the Machine Browser 
	dialog box, which contains a list of all the results in the PC Bench 
	database. To display a set of results, choose the results you want 
	and then select the Load button. Once you select the Load button, 
	PC Bench returns to the main program screen and displays the set of 
	results you chose in the results graph buttons. You can load only one 
	machine's results at a time from this window. To add another 
	comparison machine's results to the graph buttons, select Compare 
	Results again. You can display up to two results in addition to your 
	current results on the PC Bench graph buttons. You can also add 
	comparison results to the display by selecting Add Comparison System 
	from the Graph menu. (For complete information on how to view and 
	compare PC Bench results, see Chapter 4.)

+       Settings. Selecting this button displays the Machine Information 
	dialog box. From this box, you can change database information about 
	your PC that PC Bench uses to identify results you save. You can also 
	change this information by selecting Settings from the Test Suite 
	menu.

+       Help. Selecting this button opens the PC Bench help system. You can 
	also display a list of help topics by selecting the Help menu.

+       Exit. Selecting this button exits the PC Bench program screen and 
	ends the current PC Bench session. You can also exit PC Bench by 
	selecting Exit from the File menu.


Selecting tests

Before you run a PC Bench test, you must select which test suite or 
individual tests you want to run. You can select tests you want to run in two 
ways.

+       You can select a test suite or All Tests from the list of test suites 
	in the drop-down list to the right of the Run function button.

	{SYMBOL 56 \f "Wingdings"}      Click on the down arrow to the right 
		of the drop-down list. Use the scroll bar beside the list to 
		move up and down the list. Select the test suite you want by 
		clicking on its name.

	{SYMBOL 58 \f "Wingdings"}      Press the Tab key until PC Bench 
		highlights the drop-down list. (The status message will say 
		"Select group of tests to run.") Once you highlight the text 
		box, press the down-arrow key to display the list. Then, 
		highlight the test you want using your PC's arrow keys to 
		move up and down the list. Once you highlight the test you 
		want, press the Enter key to select it.

	NOTE:   With PC Bench, you can also create custom test suites 
		containing selected individual tests and groups of tests. 
		For more information, see the section "Creating and running 
		custom test suites" later in this chapter.

+       You can select individual tests or groups of tests from the Tests 
	menu.
	
	{SYMBOL 56 \f "Wingdings"}      Click on Choose Test from the Tests 
		menu. PC Bench lists the individual tests in the Test 
		Selection window. You can use the scroll bars and the Page Up 
		and Page Down keys to move up and down the list. To select a 
		test, click on the test's name. To select a group of tests, 
		click on the test group title. (The individual tests names are 
		indented under the group title.) PC Bench places a check mark 
		({SYMBOL 252 \f "Wingdings"}) beside each test you select. To 
		de-select a test, click on it again. Continue with this process 
		until you have selected all the tests you wish to run.

		To select all the tests in the list, click on the Select All 
		button. To clear your current selections, click on the Clear 
		All button. Once you have selected the test(s) you want, 
		click on the Run button.

	{SYMBOL 58 \f "Wingdings"}      Use the Tab key to highlight the 
		Tests menu option. (The status message will say "Run tests 
		quickly.") Then press the down-arrow key to display the 
		drop-down menu. Press the Enter key to select Choose Test. 
		(You can also use the Alt-t-t key combination to display this 
		window.)

		From the Test Selection window, use your PC's arrow keys and 
		the Page Up and Page Down keys to scroll through the list of 
		tests. When you highlight a test you want to run, press the 
		Enter key to select that test. To select a group of tests, 
		highlight the test group title and then press the Enter key. 
		(The individual tests names are indented under the group 
		title.) PC Bench places a check mark 
		({SYMBOL 252 \f "Wingdings"}) beside each test you select. 
		To de-select a test, highlight the test name again and press 
		the Enter key. Continue with this process until you have 
		selected all the tests you wish to run.

		To select all the tests in the list, use the Tab key to 
		highlight the Select All button and then press the Enter key. 
		To clear your current selections, highlight the Clear All 
		button and then press the Enter key. Once you have selected 
		the test(s) you want, highlight the Run button and press the 
		Enter key to start the test(s). (You can also use the 
		underlined key sequences to perform these functions.)

As PC Bench runs tests, it displays status messages at the bottom of the 
screen. When all the tests end, PC Bench displays the scores for the group 
tests on the results graph buttons. To view the individual test results for 
a group of tests, select the appropriate graph button (Processor, Video, 
Disk, or Memory). (For more information on how to view results, see the 
section "Understanding the graph buttons" in Chapter 4.)


Running the PC Bench tests

With PC Bench, you can run several kinds of tests. For example, you can run:

+       One of the PC Bench built-in test suites: DOSMark, Processor, Video, 
	or Disk. Each of these test suites contains a specific set of 
	PC Bench individual tests.

	NOTE:   A test suite is a group of tests you run at once with a 
		single command.

+       Your own custom test suites. PC Bench lets you create your own test 
	suites using the individual PC Bench tests.

+       One of the individual tests or a predefined group of these tests. You 
	can run a group of individual tests without creating a test suite.

+       All the PC Bench tests.

+       PC Bench in batch mode. You can execute PC Bench from a DOS batch 
	file using command-line options. When you run PC Bench this way, you 
	do not have to give commands to PC Bench as it runs its tests.

+       PC Bench in demonstration mode. PC Bench runs continuously in demo 
	mode until you stop it.

The individual tests report unweighted, absolute results. The DOSMark and 
subsystem scores are weighted, harmonic means of those results. Any time you 
run individual tests that are a part of a previously run DOSMark or subsystem 
test suite, PC Bench displays the absolute results for the individual test, 
recalculates the weighted scores, and updates the PC Bench display to reflect 
the new scores. For example, if you run just the 16-bit Protected Mode Large 
Mix, PC Bench will automatically recalculate your DOSMark, because it 
includes this test in the DOSMark calculation.

You can run any combination of tests during a single interactive PC Bench 
session. For example, you could run the DOSMark tests, run several individual 
tests, run a custom test suite, and then run the Disk tests.

NOTE:   When PC Bench is running a test, it disables your PC's mouse and 
	keyboard.

	Note also that if you run PC Bench from a batch file, you can run all 
	the tests at once, a single test suite, or PC Bench in demo mode. For 
	more information, see the sections "Running PC Bench in batch mode" 
	and "Running PC Bench in demonstration mode" later in this chapter.

The following sections provide details on these options for running PC Bench.


Running the DOSMark Test Suite

To run the DOSMark Test Suite, select DOSMark from the list of test suites 
next to the Run function button. Then select the Run button.

The DOSMark provides a general measure of your PC's performance. It includes 
tests from the Processor, Video, and Disk Test Suites. (For more information 
on these suites, see the next section.) 

The DOSMark tests include several disk-intensive tests that can take quite a 
while to run. Depending on your PC's configuration, it could take as long as 
45 minutes, an hour, or even longer. A top-of-the-line, correctly configured 
66-MHz 486-based PC should take less than 15 minutes to run the DOSMark Test 
Suite.

For more information on the DOSMark tests and what this score means about 
your PC, see Chapter 5.


Running other PC Bench built-in test suites

PC Bench contains three other built-in test suites you can run: the 
Processor, Video, and Disk tests. Each of these tests provides a general 
measure of its respective subsystem.

To run one of these test suites, select the appropriate test suite name from 
the list of test suites next to the Run button. Then select the Run button. 
For more information on these built-in test suites, read Chapter 5.

PC Bench also includes a separate group of Memory Tests. You run the Memory 
Tests by selecting Memory from the list of tests next to the Run function 
button and then selecting the Run button.

NOTE:   PC Bench includes the separate Memory Tests as additional performance 
	analysis tools, but the scores of these tests do not figure in the 
	calculation of the DOSMark.


Running all the PC Bench tests

To run all the PC Bench tests, select All Tests from the list of tests. Then 
select the Run button. Running all the tests can take a long time. On some 
PCs, it can take as long as two hours to run all the tests. A top-of-the-line, 
correctly configured 66-MHz 486-based PC should take less than an hour to run 
all the PC Bench tests. Once you start the tests, you do not need to interact 
with PC Bench again until the tests are complete.

While running all the PC Bench tests will produce many individual results for 
your PC, you can get a general measure of your PC's performance, without 
running all the tests. In most cases, the DOSMark will provide you with 
enough information to evaluate your PC's overall performance.

NOTE:   You should be aware of the following before you run all the PC Bench 
	tests on your PC:

	+       All disk tests will attempt to run on the same drive, because 
		PC Bench lets you set just one drive for the disk tests.

	+       To run either the Disk Throughput Tests or the Disk Mix 
		requires a minimum of 32 MB of free disk space.

	+       You will only get results for the BIOS Disk Access Tests if 
		the drive parameter in the Disk Settings section of the 
		Machine Information dialog box is set to a primary disk 
		partition (usually drive C:).

	+       The DOS File Access test will not run on a networked drive.

For a list of all the PC Bench tests, see Appendix A.


Running selected PC Bench tests

To run specific tests or sets of tests, select Choose Test from the Tests 
menu. PC Bench displays the Test Selection window.

Figure 3-3:  Selecting tests
{EMBED MSDraw   \* mergeformat|}

In the Test Selection window, select the test or tests you wish to run. To 
select a group of tests, select the test group title (such as Processor 
Tests). The list indents the individual tests names under each group title 
name. When you select a test, PC Bench marks that test with a check mark 
({SYMBOL 252 \f "Wingdings"}). To de-select a test, simply choose that test 
again. PC Bench removes the check mark. To select all the tests in the list, 
select the Select All button. To de-select all of your current selections, 
select the Clear All button. To exit the window, select the Close button.

Once you have selected the test or tests you wish to run, select the Run 
button at the bottom of the dialog box. PC Bench then runs the tests you 
marked.


Running PC Bench in batch mode

You can execute PC Bench from a DOS batch file using command-line options. A 
batch file contains a series of commands DOS will perform automatically once 
you execute the file. When you run PC Bench this way, you do not have to 
interact with PC Bench as it runs its tests, and PC Bench automatically logs 
the test results in its database.

If you have more than one drive on your PC, your batch file should first 
specify the letter of the drive that contains the directory where you loaded 
PC Bench. Next, the batch file should change to that directory (CD) (probably 
\PCBENCH) and execute the BENCH command with the /B option and any other 
options you choose.

Other options you can use on the command line in the batch file are: /Ffile, 
/M, /Ofile, and /P. For a description of these command-line options, see the 
section "Using PC Bench command-line switches" earlier in this chapter.


Sample batch files

The following paragraphs contain sample batch files. You must set the PIN 
environment variable and all the VARIANT environment variables from your 
batch file. (For more information on the DOS environment variables you can 
use with PC Bench, see the section "Using PC Bench command-line switches" 
earlier in this chapter.)

1.      The following batch file changes to the \PCBENCH directory, sets the 
	necessary environment variables, starts PC Bench in batch mode from 
	your PC's hard disk, disables system information logging, and saves 
	the test results in the file PC1.CSV.

	CD \PCBENCH
	SET PIN=PC25
	SET VARIANT1=1
	SET VARIANT2=0
	SET VARIANT3=0
	SET VARIANT4=0
	SET VARIANT5=0
	BENCH /B /M /OPC1.CSV

2.      The following batch file changes to drive D:, changes to the \PCBENCH 
	directory, sets the necessary environment variables, starts PC Bench 
	in batch mode, and runs the tests listed in the test suite file 
	TEST1.

	D:
	CD \PCBENCH
	SET PIN=PC25
	SET VARIANT1=2
	SET VARIANT2=0
	SET VARIANT3=0
	SET VARIANT4=0
	SET VARIANT5=0
	BENCH /B /FTEST1.STE

3.      The following batch file changes to the \PCBENCH directory, sets the 
	necessary environment variables, runs PC Bench in batch mode from 
	your PC's hard disk, prevents PC Bench from automatically determining 
	your processor type, and saves the test results in the file PC16.CSV.

	CD \PCBENCH
	SET PIN=PC3
	SET VARIANT1=1
	SET VARIANT2=2
	SET VARIANT3=0
	SET VARIANT4=0
	SET VARIANT5=0
	BENCH /B /P /OPC16.CSV


Running PC Bench in demonstration mode

Once you install PC Bench, you can run it in demonstration mode. To execute 
PC Bench in demo mode, select Run Demo from the File menu. To halt demo mode 
press the Esc key.

NOTE:   When you execute PC Bench in demo mode, it runs the current test 
	suite or set of tests listed in the drop-down list beside the Run 
	function button. Thus, you can select which test you want PC Bench to 
	run in its demo by changing the current test suite.

Once PC Bench completes a test run in demo mode, it displays the test results 
for that run briefly on the results graph buttons and then runs the same test 
suite again. PC Bench runs continually in demo mode until you stop it.


Running the tests included as utilities

PC Bench includes a set of tests you can use as utilities. These tests are 
the Battery Rundown, the Time the Timer, the VGA Compatibility, and the 
EGA/VGA Monitor Quality tests. You can not execute these tests from within 
PC Bench. Instead, you must execute them from the DOS command line.

The tests PC Bench includes as utilities are:

+       The Battery Rundown Test. To run this test, execute the following 
	command from the directory where you installed PC Bench:

	RUNDOWN.EXE

+       The Time the Timer Test. To run this test, execute the following 
	command from the directory where you installed PC Bench:

	TIMER.EXE

+       The VGA Compatibility Test. To run this test, execute the following 
	command from the directory where you installed PC Bench:

	VGAHDW.EXE

+       The EGA/VGA Monitor Quality Test. To run this test, execute the 
	following command from the directory where you installed PC Bench:

	EGAVGA.EXE

For more information on these tests, see Appendix A.


Saving your test results

To commit the current test results to PC Bench's database, select Save from 
the File menu. PC Bench automatically updates its database with the results 
information for your current session and stores your PC's configuration 
information with these results.

NOTE:   Because PC Bench logs your system configuration information each time 
	you save results, it may take a while to complete the save.

If you exit PC Bench without saving your current results, PC Bench displays a 
warning message to remind you that you have not saved your current results. 
You can select Save, Delete, or Cancel in this warning box.


Creating and running custom test suites

A test suite is a set of tests you execute with a single command. PC Bench 
lets you create your own test suites. This set of tests can consist of any 
combination of the PC Bench tests you want. You can place both individual 
tests and test groups in your test suite. You can also import previously 
created test suites. The following sections describe how to create and edit 
test suites and how to use previously created test suites.


Creating a custom test suite

To create a custom test suite:

1.      Select a test suite from the list of test suites beside the Run 
	button that is similar to the type of test suite you wish to create.

	If there is not a test suite similar to the type you want to create, 
	you may wish to select All Tests. At least this way, you can begin 
	with all the tests selected and de-select the tests you do not want 
	in your test suite.

2.      Select Edit from the Test Suites menu.

	PC Bench displays the Test Selection window.

	Figure 3-4:  The Test Selection window
	{EMBED MSDraw   \* mergeformat|}        
	
	This window displays a list of the PC Bench tests you can use in your 
	custom test suite. PC Bench marks the tests defined in the currently 
	selected test suite with a check mark ({SYMBOL 252 \f "Wingdings"}).

3.      Select the test or tests you wish to use in your custom test suite.

	To de-select all the tests currently marked, select the Clear All 
	button. To select all the tests in the list, select the Select All 
	button. To de-select a single test, choose the test's name again to 
	remove the check mark.

4.      Select the Close button once you have defined the tests for your test 
	custom suite.

	The list of test suites beside the Run function button will display 
	"<current>," indicating that there is a new test suite you have not 
	yet saved.

	NOTE:   You can run the tests you selected at this point without 
		saving the test suite by selecting the Run function button. 
		However, if you do not save the test suite, it will not be 
		available for future PC Bench sessions.

5.      Select Save from the Test Suites menu to save the tests you defined 
	in the Test Selection window.

	NOTE:   When you save your custom test suite, PC Bench also saves 
		your disk test settings. If you wish to save your test suite 
		with different disk test settings, either select the Settings 
		option from the Test Suite menu or use the Settings function 
		button before you save the results. (Both options will open 
		the Machine Information dialog box.)

6.      Enter a name for your new test suite in the Save File As window and 
	then select the Save button.

	The name you enter here should be a legal DOS file name with no file 
	name extension. PC Bench automatically appends the extension .STE to 
	the file name you enter here.

PC Bench saves the information you entered and adds that test suite file name 
to the list of test suites in the drop-down list beside the Run button. You 
can use your custom test suite during your current PC Bench session or in a 
future PC Bench session. To run the test suite, simply select the test suite 
name from the list and then select the Run button.


Editing a custom test suite

You can change the tests that comprise an existing test suite by adding or 
deleting tests. To do this:

1.      Select the test suite you wish to edit from the list of test suites 
	beside the Run function button.

2.      Select Edit from the Test Suites menu.

	PC Bench displays the Test Selection window. This window lists all 
	the available PC Bench tests. The tests that make up the test suite 
	you selected are marked with a check mark 
	({SYMBOL 252 \f "Wingdings"}).

3.      Select or de-select tests. Once you have the test suite defined 
	the way you want, select the Close button.

4.      Select Save from the Test Suites menu to save the edited test suite.

5.      Overwrite the old test suite in the Save File As dialog box by 
	selecting that test suite file name.

PC Bench automatically updates the test suite file. The next time you run the 
test suite, it will run the new set of tests you selected.


Using previously created test suites

To use a previously created test suite, simply copy the file you want into 
the PC Bench directory on your PC.

If you are getting the test suite file from another PC, you can copy the file 
by putting it on a diskette, or if you are on a network, you can copy the 
file across the network. Make sure the file name has a .STE extension. Once 
the test suite file is in the PC Bench directory on your PC, the test suite 
name will appear in the list of tests the next time you run PC Bench.

NOTE:   Be aware that if a test suite file from another PC uses a disk drive 
	not available on your PC, PC Bench will use your PC's current drive 
	for the test.


What's next

Now that you know how to run the PC Bench tests, Chapter 4 tells you how to 
view and compare results.

If you would like information on the different test suites that come with 
PC Bench or if you would like a better understanding of what the tests do, 
read Chapter 5.

End of Chapter



Chapter 4
Working with PC Bench's Results


This chapter explains how you can work with PC Bench results. This chapter 
contains sections that tell you how to: 

+       Read and understand the test results in the results graph buttons in 
	the PC Bench program screen.

+       Compare results from different test runs on your PC or other PCs.

+       Print the results for a test run.

+       Save test results to the PC Bench database.

As you read this chapter, you should understand that the current baseline 
system is always the set of test results to which PC Bench compares all other 
test results. PC Bench lets you select which set of results you want to use 
as your baseline system. The baseline system can be your current set of 
results or a set of results you saved from a different run of the tests. By 
default, PC Bench sets the baseline system's results to 100% and displays all 
other sets of results as a percentage of that PC's results. For example, 
since the baseline system is 100%, a slower set of results appears as less 
than 100% and a faster set of results appears as higher than 100%. (For 
complete information on setting your baseline system, see the section 
"Selecting a baseline system" later in this chapter.)


Understanding the graph buttons

When PC Bench completes a test run, it displays its results in graph format 
on the graph buttons in the PC Bench Results portion of the program screen.

Figure 4-1:  PC Bench Results
{EMBED MSDraw   \* mergeformat|}        

NOTE:   If you have not yet run any PC Bench tests or if you have not run 
	enough tests to compute the DOSMark or subsystem scores, the graph 
	buttons for the DOSMark and subsystem scores will read "(Insufficient 
	Data to Compute Score)". In addition, the Memory Tests graph button 
	will read "(No Results)" when there are no Memory Tests results.

	Note also that if you exited a previous session of PC Bench while 
	PC Bench was displaying a set of results, PC Bench displays that same 
	set of results the next time you start a session.

Each graph for the Processor, Video, and Disk scores is actually a large 
button. There is also a separate button for any Memory Tests results. 
(PC Bench includes the separate Memory Tests as additional performance 
analysis tools, but the scores of these tests do not figure in the 
calculation of the DOSMark.) When you select a graph button, PC Bench then 
displays the bar graph results for the individual tests associated with the 
button you selected.

NOTE:   If you have a color monitor, the titles for the DOSMark, Processor, 
	Video, and Disk scores are blue. The blue titles indicate which tests 
	PC Bench includes in the DOSMark tests. If you do not have a color 
	monitor, the titles for these graph buttons will be slightly lighter 
	than the title for the Memory Tests button, since the Memory Tests 
	results are not part of the DOSMark.

Each graph button for tests included in the DOSMark supplies the following 
information:

+       The name of the test PC Bench ran.

+       The name of your PC as set in the Machine Information dialog box.

	NOTE:   You can change the graph so the name it displays for the PC 
		is either the Machine ID, the Machine Name, or the Variants 
		set when you ran the test. To change this view, select Show 
		Name, Show ID, or Show Variants from the Graph menu.

+       The absolute score your PC received on the test. (If you use the 
	Compare Results feature, you will also see a score for each set of 
	results you choose to add to the graph.) The absolute score is the 
	number to the left of the bar graph.

+       A bar graph displaying the score as a percentage. If you are 
	displaying only one set of results, those results are set to 100%. If 
	you are displaying more than one set of results and if you have not 
	selected a baseline system, PC Bench sets the largest set of results 
	for any given test to 100%. If you have chosen a baseline system, the 
	bar graph score for each set of results is relative to the bar graph 
	score for the baseline system. By default, the set of results for the 
	first tests you run become the baseline system. (For more information 
	on the baseline system, see the section "Selecting a baseline system" 
	later in this chapter.)

NOTE:   To display the bar graph according to the absolute numbers instead of 
	a percentage, select Value from the Graph menu.

To see results for the individual tests PC Bench ran as part of a group of 
tests, select the button for the appropriate test group. For example, to view 
the results for the Processor Tests, select the graph button for the 
Processor Score.

	{SYMBOL 56 \f "Wingdings"}      Click directly on the graph button 
		for the appropriate group of tests whose results you want to 
		view. For example, if you want to view your PC's scores for 
		the individual video tests, click on the Video Score graph 
		button.

	{SYMBOL 58 \f "Wingdings"}      Use the Tab key to highlight the 
		appropriate graph button and then press the Enter key to 
		select that graph's button. For example, if you want to view 
		your PC's scores for the individual processor tests, use the 
		Tab key to highlight the Processor Score graph button and 
		then press the Enter key. You can also use the key 
		combination Alt-v to move directly to the graph buttons and 
		to toggle between the graph buttons and the graphs for the 
		individual tests.

Once you select the appropriate graph button, PC Bench displays that group's 
individual tests results. If there is more than one page of results, you will 
see a Next Page button at the bottom of the screen. You can use the Next Page 
(and Previous Page) buttons to page through the results screens for the 
individual tests.

Figure 4-2: The Next Page and Previous Page buttons
{EMBED MSDraw   \* mergeformat|}

For more information on what the PC Bench results mean about your PC's 
performance, see Chapter 5. For more information on how to use the graph 
buttons to compare results, see the next section.


Comparing results

To compare different PC Bench results, select the Compare Results function 
button from the PC Bench program screen. PC Bench displays the Machine 
Browser window.

Figure 4-3:  The Machine Browser window
{EMBED MSDraw   \* mergeformat|}

From this window, you can select which set of results you want to display in 
the PC Bench Results portion of the screen. PC Bench lists all the results 
available in its database in this window.

NOTE:   If you run PC Bench repeatedly on the same machine, the Machine ID or 
	Variants for each test run must be unique. PC Bench automatically 
	makes each test run unique by changing the Variant fields. You can, 
	however, modify these fields from the Machine Information dialog box 
	and use your own naming scheme or convention. (For more information, 
	see "Changing your PC's test settings" in Chapter 3.)

	For example, if you are testing a PC with different software disk 
	caches, you could use one variant field to identify the different 
	types of disk cache and another variant field to identify the size of 
	the disk cache.

The following sections explain how to compare, view, and manipulate the 
results PC Bench displays.


Selecting a baseline system

You can compare all the results to one set of results. PC Bench considers 
this set of results the baseline system. PC Bench lets you choose which set 
of results to use as your baseline system. PC Bench sets those results to 
100% and displays all other results in the graphs according to how they 
compare with this baseline system's results. For example, the bar graph for a 
set of results that are faster than the baseline system will be more than 
100%, while the bar graph for a set of results that are slower than the 
baseline system will be less than 100%.

To specify which set of results you want PC Bench to use as the baseline 
system or to change the current baseline system, select Select Baseline 
System from the Graph menu.

Figure 4-4:  Selecting the baseline system
{EMBED MSDraw   \* mergeformat|}

PC Bench displays a submenu listing the machine names for your current test 
run and for all comparison results you have added to the display. (For more 
information on how to add a set of results to the display, see the next 
section, "Adding a set of results to the display.")

At this point, you can either select the machine name whose set of results 
you wish to use as your baseline system, or you can select "Largest."

{SYMBOL 56 \f "Wingdings"}      Click on Graph in the menu bar and then click 
	on Select Baseline System. When PC Bench displays the list of 
	available machine names in the cascading menu, select the one you 
	want by clicking on it.

{SYMBOL 58 \f "Wingdings"}      Use the Alt-g-s key combination to display 
	the cascading menu. Use your PC's up and down arrow keys to highlight 
	the machine name you want and then press the Enter key to select it.

If you select "Largest," PC Bench sets the highest results for any given 
individual test as the baseline system for that test. Using "Largest" is a 
quick and easy way to see which results for which tests are the highest.

NOTE:   Because PC Bench lets you select the baseline system, you can change 
	the baseline system to a different set of results at any time. You 
	may wish to change the baseline system to compare results for 
	different tests.

You can also display the results graphs relative to the absolute numbers 
instead of a percentage by selecting Value from the Graph menu.

NOTE:   To see the differences in machine results better, you can use the 
	Zoom In and Zoom Out options in the Graph menu. The zoom feature is 
	useful when viewing absolute numbers.

	When you zoom in, PC Bench displays a more detailed view of the bar 
	graph results. For example, if PC Bench is displaying the results in 
	a percentage and the scale numbers are 100% and 200%, when you zoom 
	in, PC Bench displays the scale numbers as 50% and 100%. By zooming 
	in, you can read the bar graph results more easily when two tests 
	have results that are very close.

	When you zoom out, PC Bench compresses the size of the bar graph so 
	more of the graph displays in the same amount of space. For example, 
	if PC Bench is displaying the results in a percentage and the scale 
	numbers are 100% and 200%, when you zoom out PC Bench displays the 
	scale numbers as 250% and 500%. 
	

Adding a set of results to the display

You can add a set of results to the display for comparison with your current 
results or with other results. You can display up to three sets of results: 
your current results and two other comparison systems. To add a comparison 
system:

1.      Select either the Compare Results function button or Add Comparison 
	System from the Graph menu. (Selecting either will give you the same 
	dialog box.)

	{SYMBOL 56 \f "Wingdings"}      Click on the Compare Results function 
		button or click on Add Comparison System from the Graph menu.

	{SYMBOL 58 \f "Wingdings"}      Use either the Alt-c or the Alt-g-a 
		key combination.

	PC Bench displays the Machine Browser dialog box, which lists all the 
	results available in the database.

2.      Select the results you want to add to the results graph buttons.

	{SYMBOL 56 \f "Wingdings"}      Click on the machine name and then 
		click on the Load button.

	{SYMBOL 58 \f "Wingdings"}      Use your PC's up and down arrow keys 
		to highlight the machine name and then use the Alt-l key 
		combination to load that PC's results. (You can also use the 
		Tab key to highlight the Load button and then press the Enter 
		key to select it.)

PC Bench adds the results you selected to the results displayed on the graph 
buttons.


Importing PC Bench results

If you have access to PC Bench results for other PCs, you can import those 
results into your PC Bench database and then use them as comparison points 
with your PC's results. PC Bench includes batch files you can use to set up 
an import database in your PC Bench directory. Before you execute these batch 
files, you should put the results you want to import in a directory 
accessible from your PC.

NOTE:   You can import only one PC Bench database of results at a time.

To import results:

1.      Execute the following command from the directory where you installed 
	PC Bench (probably \PCBENCH):

	SETUPIMP DIR

	where DIR is the pathname for the database files you wish to import.

	NOTE:   The pathname should end with a backslash (\). For example, if 
		the database files you want to import are in C:\PC1\RESULTS, 
		you would enter the following command:

		SETUPIMP C:\PC1\RESULTS\
	
		Or if the import database files are on a floppy diskette, and 
		your PC's floppy drive is drive A, you would enter the 
		following command:

		SETUPIMP A:\
	
		The directory you specify in the pathname should contain the 
		following PC Bench database files:

		MACHINE_.DBF
		RESULT_.DBF
		SYSINFO_.DBF
		MACHINE_.MDX
		RESULT_.MDX
		SYSINFO_.MDX

		NOTE:   DO NOT copy the files you want to import directly 
			into your PC Bench directory. If you do, you will 
			overwrite your database.

		PC Bench automatically renames these files as it copies them 
		to your PC Bench directory so that you do not overwrite your 
		own existing database files.

2.      Start PC Bench and select Import from the File menu.

	{SYMBOL 56 \f "Wingdings"}      Click on Import from the File menu.

	{SYMBOL 58 \f "Wingdings"}      Use the Alt-f-i key combination.

	NOTE:   If the import database is very large, you will need to free 
		as much conventional memory as possible to load the list of 
		results into the Import Machine Browser window. If you get a 
		memory error while trying to import results, exit PC Bench, 
		free up memory, and try importing the results again.

	PC Bench displays the Import Machine Browser window showing you the 
	different sets of results from the other PC.

	Figure 4-5:  The Import Machine Browser window
	{EMBED MSDraw   \* mergeformat|}

3.      Select the set of results you wish to import into the PC Bench 
	database on your PC.

	{SYMBOL 56 \f "Wingdings"}      Click on the set of results and then 
		click on the Import button.

	{SYMBOL 58 \f "Wingdings"}      Use your PC's up and down arrow keys 
		to highlight the set of results, press the Enter key to mark 
		those results, and then use the Alt-i key combination to 
		import that PC's results. You can also use the Tab key to 
		highlight the Import button and then press the Enter key.

	You can import all the results in the other PC's database by 
	selecting the All button at the bottom of the screen.

	NOTE:   Depending on the number of results in the database you are 
		importing, it may take a while to import them.

You can repeat the above steps for each database of results you want to 
import.

Once you import the results you want, you can save disk space by deleting the 
imported database files. To delete these files, execute the following command 
from the PC Bench directory:

CLEARIMP

This command deletes the database files created by the SETUPIMP command.

Removing a set of results from the displayed results
To remove a set of comparison results from the PC Bench Results portion of 
the screen:

1.      Select Remove Comparison System from the Graph menu.

	{SYMBOL 56 \f "Wingdings"}      Click on Graph in the menu bar and 
		then click on Remove Comparison System.

	{SYMBOL 58 \f "Wingdings"}      Use the Alt-g-r key combination.

	PC Bench displays a cascading menu that lists all the results 
	PC Bench is currently displaying.

	Figure 4-6:  Removing a set of results from the display
	{EMBED MSDraw   \* mergeformat|}

2.      Select the set of results you want to remove from the display.

	{SYMBOL 56 \f "Wingdings"}      Click on the machine name, 
		machine ID, or variants.
		
	{SYMBOL 58 \f "Wingdings"}      Use your PC's up and down arrow keys 
		to highlight the machine name, machine ID, or variants, and 
		then press the Enter key.

	NOTE:   The type of display you have chosen from the Graph menu (Show 
		Name, Show ID, or Show Variants) determines what type of 
		identifier appears in this list.

PC Bench removes that set of results from the graph buttons and re-displays 
the results. It does not, however, delete these results from the PC Bench 
database. You can display these results again if you choose to do so.


Deleting results from the PC Bench database

To delete results from the PC Bench database permanently:

1.      Select Delete Results from the File menu.

	{SYMBOL 56 \f "Wingdings"}      Click on File in the menu bar and 
		then click on Delete Results.

	{SYMBOL 58 \f "Wingdings"}      Use the Alt-f-d key combination.

	PC Bench displays the Machine Browser window. In it are the sets of 
	results in the database that do not currently appear in the 
	PC Bench Results portion of the screen.

	NOTE:   You cannot delete a set of results you are viewing from the 
		PC Bench database. To delete a set of results you are 
		viewing, first remove that set of results from the display 
		using the Remove Comparison System option from the Graph menu.

2.      Select the set of results you want to remove from the database.

	{SYMBOL 56 \f "Wingdings"}      Click on the set of results and then 
		click on the Delete button.

	{SYMBOL 58 \f "Wingdings"}      Use your PC's up and down arrow keys 
		to highlight the set of results, press the Enter key to mark 
		those results, and then use the Alt-d key combination to 
		delete the results. You can also use the Tab key to 
		highlight the Delete button and then press the Enter key.

PC Bench deletes that set of results from the database.


Viewing your PC's configuration information

Because your PC's configuration affects your PC Bench scores, PC Bench 
automatically logs configuration information when you save results. To view 
the configuration information, select the System Info function button from 
the PC Bench program screen.

NOTE:   You can also view your PC's configuration information by selecting 
	the Display System Info option from the File menu (Alt-f-i).

PC Bench displays your PC's configuration information in the System 
Information dialog box.


Figure 4-7:  The System Information dialog box
{EMBED MSDraw   \* mergeformat|}

NOTE:   When you select the System Info function button for the first time 
	during your current PC Bench session, PC Bench must gather all the 
	information about your PC. Thus, it may take PC Bench a while before 
	it displays the System Information window.

The System Information window contains data on the configuration that 
produced each set of results currently loaded in the PC Bench Results 
portion of the screen. You can also view the CONFIG.SYS and AUTOEXEC.BAT 
files for any of these sets of results.

NOTE:   When you view information in this window, you can use the scroll bar 
	on the right to page up and down through the information in each 
	screen.

The following list explains briefly the information you can get from each of 
the buttons in this window.

+       The Hardware button. By default, when you open the System 
	Information window, PC Bench displays your PC's hardware information. 
	This screen tells you information about your PC's microprocessor type 
	and CPU clock speed, plus other information about your PC's hardware.

+       The System button. When you select this button, PC Bench displays 
	information about your PC's operating system type and version as well 
	as information about any disk caching software you have loaded on 
	your system. You can also find the current date and time in this 
	screen plus other information about your PC's software.

+       The Memory button. Selecting this button displays information about 
	your PC's memory subsystem. For example, this screen tells you the 
	amount (in kilobytes) of your PC's conventional, expanded, and 
	extended memory.

+       The Disk button. When you select this button, PC Bench displays 
	information on your PC's disk subsystem. For example, in this screen 
	you find the size (in bytes) of your PC's fixed disk. The screen also 
	provides details on the disk partitions of the fixed disk (such as 
	the size, in megabytes, and type of partitions).

+       The Config.sys and Autoexec.bat buttons. These buttons display your 
	PC's current CONFIG.SYS and AUTOEXEC.BAT files. (To edit these files, 
	you must exit PC Bench and use a DOS file editor.)


Printing results

Once you have added comparison results to the PC Bench Results portion of the 
screen, you can print those results to a file or to your local printer.

To print test results, select Print from the File menu. PC Bench displays the 
Print Results dialog box.

Figure 4-8:  Printing results
{EMBED MSDraw   \* mergeformat|}

This box lists all the sets of results currently loaded in the PC Bench 
Results portion of the screen. You can print only one set of results at a 
time. To print a set of results:

1.      Select which set of results you want to print from the list at the 
	top of the Print Results dialog box. An "X" appears beside the 
	results you select.

	{SYMBOL 56 \f "Wingdings"}      Click on the box beside the set of 
		results.

	{SYMBOL 58 \f "Wingdings"}      Use the Tab key to highlight the box 
		beside the set of results and then press the Enter key.

2.      Mark whether you want to print the set of results to a file or to a 
	local device (such as a printer). An "X" marks your selection.

	The list of devices in the Device drop-down list are PRN, LPT1, LPT2, 
	COM1, and COM2. Which device you select depends on which of your PC's 
	ports the printer is connected to.

	To print the results to a file, specify a file name in the box to the 
	right of File. The file name must be a legal DOS file name. PC Bench 
	saves the file in the current directory.

	To enter this information:

	{SYMBOL 56 \f "Wingdings"}      Click on the box beside either Device 
		or File. You can display a list of devices by clicking on the 
		down arrow to the right of the drop-down list beside Device.

	{SYMBOL 58 \f "Wingdings"}      Use the Tab key to highlight the 
		drop-down list to the right of either Device or File and 
		then press the Enter key. To display a list of devices, use 
		your PC's right arrow key to move the focus to the list and 
		then press the down arrow key to display the list.

3.      Select the Print button to print the set of results.

If you choose to print your PC's results to a file, you have two choices for 
your file's format: ASCII text (.TXT) or comma separated values (.CSV) 
format. Enter a legal DOS file name without an extension in the text box 
beside the File option. PC Bench automatically adds the correct extension to 
the file name you enter (either .TXT for ASCII text or .CSV for comma 
separated value).

NOTE:   If you print your PC's results to a file using CSV format, you can 
	read that CSV file into many database or spreadsheet programs to view 
	or reformat your results.

	Note also that when you print a set of results, the print output and 
	the file will contain a "Not Run" entry for each individual test 
	PC Bench did not run. For example, if you print a set of results that 
	contains results for only the Processor tests, PC Bench lists all the 
	other tests as "Not Run."


What's next

Now that you understand how to view and compare PC Bench results, Chapter 5 
provides details on what operations the PC Bench tests perform and what your 
PC's scores mean about system performance.

Chapter 6 lists some of the factors that might influence your PC's results 
and gives tips on how to configure your PC to obtain the best possible 
PC Bench results.

If you would like to know more about the concepts behind PC Bench's 
development, see Chapter 7.


End of Chapter



Chapter 5
Understanding PC Bench's Results


This chapter discusses the different PC Bench tests and how you can use their 
results. After reading this chapter, you should know what the PC Bench 
results mean about your PC's performance.

For a list of the tests that come with PC Bench and a complete description of 
each test, see Appendix A.


The PC Bench scores

Which PC Bench tests you should choose to run depends on how you want to test 
your PC. The following sections describe the different types of PC Bench 
tests and how these tests measure your PC's performance. (For information on 
how to run the PC Bench tests, see the section "Running the PC Bench tests" 
in Chapter 3.)


The DOSMark

The DOSMark provides an overall measure of your PC's performance. It consists 
of tests that reflect how your PC performs when running DOS applications.

To produce a DOSMark, PC Bench runs the tests necessary to produce the 
Processor, Video, and Disk scores and then calculates the DOSMark as a 
weighted harmonic mean of those three scores. You can get an overall measure 
of your PC's performance, and you can also consult the individual test 
results that make up the DOSMark to understand better how your PC's 
subsystems perform.

NOTE:   If you have a color monitor, you will notice the titles for the 
	DOSMark, Processor, Video, and Disk graph buttons are in blue. The 
	blue titles indicate which tests PC Bench includes in the DOSMark. If 
	you do not have a color monitor, the titles for these graph buttons 
	will be slightly lighter than the title for the Memory Tests button.

PC Bench uses tests based on application profiling to gauge your PC's 
performance, so the DOSMark represents a very close portrayal of application 
performance without forcing you to run a large set of applications.

NOTE:   Because PC Bench's profiling used a set of representative 
	applications, individual applications you use on your PC may not 
	reflect that same performance behavior as the DOSMark. For example, 
	if your PC has a DOSMark twice that of another PC, it does not mean 
	every individual application will run twice as fast on your PC as on 
	the other PC. However, your overall DOS application performance will 
	be twice as fast.

You can use the DOSMark to compare your PC with other DOS-compatible PCs. You 
can also use the DOSMark to compare the effect of hardware or software 
changes or upgrades on your PC's overall performance.


The Processor Score

The Processor Score is a measure of how your PC's processor and memory 
subsystems perform.

The core of PC Bench's Processor Score is the 16-bit Protected Mode Large 
Instruction Mix. This mix reflects the larger working-set sizes that today's 
applications use. The 16-bit Protected Mode Large Instruction Mix stresses a 
PC's bus, CPU cache, CPU cache interface, external cache, main memory, and 
CPU instruction set in the same way as top-selling DOS applications.

The Processor Test Suite also includes a test of the performance of your PC's 
floating point unit (FPU).

When you run the Processor Test Suite, PC Bench takes the results for the CPU 
and FPU tests and calculates a weighted harmonic mean to produce the overall 
Processor Score. By weighting these results, PC Bench can reflect the 
relative importance of the CPU and FPU to profiled applications.

NOTE:   PC Bench includes a Math Coprocessor test that requires an FPU. If 
	your PC does not have an FPU, PC Bench uses FPU-emulation for this 
	particular test. Results with FPU-emulation will be significantly 
	slower than those with a hardware FPU.


The Video Score

The Video Score provides a measure of your PC's VGA video subsystem 
performance.

The Video Score is a combination of several different individual video tests. 
PC Bench uses low-level video tests to measure how your PC's VGA subsystem 
performs when running applications under DOS.

When you run the Video Test Suite, PC Bench takes the results for the 
individual video tests in the test suite and calculates a weighted harmonic 
mean of those results. These individual tests include 8- and 16-bit VGA write 
operations involving both text and graphics. The Video results give you a 
breakdown of the different types of DOS video calls that applications make.


The Disk Score

The Disk Score lets you gauge the performance of your PC's disk subsystem, 
which includes the hard disk, disk controller, and any hardware or software 
disk caches. PC Bench 8.0 uses a new disk testing methodology, its Disk Mix.

ZDBOp based the Disk Mix on the disk usage of a set of profiled, leading DOS 
applications. The Disk Mix works with your PC's disk subsystem the way actual 
applications do.

The Disk Mix performs a range of disk activities involving a large set of 
files. It works with over 100 files spread across six directories. The Disk 
Mix works with multiple files concurrently, creates and deletes files, reads 
and writes file data in different sizes, intersperses write and read 
operations, and generally uses the disk subsystem the way applications do. 
The Disk Mix works with a PC's disk subsystem via the standard DOS Int 21H 
interface, so you can use the Disk Score as an accurate measure regardless of 
whether you are using a software or hardware disk cache.


The Memory Tests results

The Memory Tests measure the speed with which your PC can access various 
types of memory. In general, the Memory Tests results reflect the speed and 
architecture of your PC's memory, memory caches (if any), and CPU type. These 
tests use a series of low-level memory read and write operations to 
conventional, extended, graphic, text, and BIOS memory.

PC Bench does not weight the Memory test results as part of its DOSMark. 
These tests are, however, are useful tools for evaluating memory subsystem 
speed.

NOTE    Although PC Bench includes a separate set of Memory Tests, ZDBOp 
	recommends you use the Processor Test Suite to measure memory 
	performance in conjunction with the processor. ZDBOp includes the 
	Memory Tests with PC Bench version 8.0 as additional informational 
	and diagnostic tools.


Other PC Bench tests

The four PC Bench overall weighted scores (the DOSMark and the Processor, 
Video, and Disk Scores) use only a subset of the tests available with 
PC Bench. PC Bench contains a large number of individual tests that measure 
the performance of different aspects of your PC's subsystems.

NOTE:   For a complete list of the individual tests included with PC Bench 
	and a brief description of each test, see Appendix A. For 
	instructions on how to run a PC Bench test, see the section "Running 
	the PC Bench tests" in Chapter 3.

You can combine the individual tests in any way you want and run them. You 
can also create your own test suite using these individual tests.

Some of the individual PC Bench tests are:

+       The Disk Throughput Tests. If you want to measure your PC's disk 
	subsystem performance when manipulating files with specific sizes in 
	specific block sizes, use the Disk Throughput Tests. You should note, 
	however, that due to the low-level approach of these tests, they may 
	not give the best reflection of a PC's disk performance when you are 
	using a disk cache.

+       The Disk Access Tests. The BIOS Disk Seek Tests (sequential and 
	random) measure mechanical track-to-track disk drive access times. 
	The tests use the BIOS Int 13H. You cannot use this test to check 
	floppy diskettes or mass storage media that do not support the 
	standard Int 13H interface. The tests involve 1,000 sequential seeks 
	(alternating between cylinders 0 and 11) and 1,000 random seeks. 
	Caching disk controllers and certain SCSI adapters that do not force 
	the disk mechanism to move for disk seeks will yield unrealistically 
	fast times on this test.

	The DOS Disk Access Test measures the time necessary to perform 1,000 
	read requests at random locations on the disk using the DOS Int 25H. 
	This test should work with any device that DOS recognizes as a disk 
	except CD-ROM or network drives. The test reports the total time (in 
	seconds) required to complete the requests.


Why results differ

If you run the same PC Bench test on your PC multiple times and use proper 
testing procedures, you will get essentially the same results every time. 
(The margin of error in the PC Bench results is about one percent.) Proper 
testing procedures include defragmenting your PC's disk and rebooting your 
PC before each test run.

If you modify any of your PC's components between test runs or deviate from 
proper testing procedures, you probably will see a difference in the results. 
A fragmented disk, for example, can affect the PC Bench results for your PC.

Naturally, there are several hardware and software components that can affect 
your PC's performance. For a list of some of the factors that can influence 
PC Bench results, see Chapter 6.


Publishing PC Bench results

The PC Bench License Agreement requires that you include certain information 
about your PC when you publish any PC Bench results. Thus, if you publish 
your PC's results, you would need to say something similar to:

Ziff-Davis' PC Bench (TM) 8.0 produced a DOSMark (TM) score of X on WXY 
PC Corp. Model 466 with a 66-MHz Intel (R) 486DX2-66 CPU, 8MB of RAM, 64 KB 
RAM cache, 200 MB hard disk, IDE controller, and no hardware disk cache. The 
test PC used XYZ Corp. XYZ Video adapter with a refresh rate of 72 Hz. The 
test PC was running MS-DOS (R) 5.0, SMARTDRV version X.Y with 2 MB cache, 
and disk compression utility ABC version 1.0 enabled. All products used in 
the test were shipping versions available to the general public, and 
the test and its results were not verified by Ziff-Davis.

PC Bench (TM) and DOSMark (TM) are trademarks of Ziff-Davis Publishing 
Company, L.P.

To find out exactly what you must include when you publish PC Bench results, 
consult the PC Bench License Agreement at the front of this manual and in the 
README.TXT file. You can also view the license agreement from the Help menu 
in the PC Bench program screen.


Using PC Bench results

You can run PC Bench as many times as you wish on your PC and save the 
results for each test run in the PC Bench database. You can then use 
PC Bench results to compare different configurations on the same PC or to 
compare the performance of different PCs. For example, you can:

+       Run PC Bench multiple times on a single PC with multiple 
	configurations and then compare the results for each test run.

+       Compare PC Bench results for your PC with the results of other PCs 
	using results published in various Ziff-Davis publications. (For a 
	list of these publications, see Chapter 8.)

+       Run PC Bench on different machines within your organization to 
	compare their performance. Running PC Bench on different machines 
	may be helpful when your organization has more than one type of PC or 
	has PCs with different configurations.

If you have access to PC Bench results for different PCs, you can import 
those results into the PC Bench database on one PC for comparison. (For more 
information, see the section "Importing PC Bench results" in Chapter 4.) You 
may also be able to download PC Bench results from ZiffNet, or you can use 
the small set of comparison results that come with PC Bench.

NOTE:   PC Bench includes sets of sample results in its database you can use 
	as comparison systems.

To view PC Bench results, use the Compare Results function button from the 
PC Bench main screen. For more information, see Chapter 4.


Using PC Bench with other benchmarks

Although PC Bench measures the speed of the processor, memory, disk, and 
video subsystems using the same operations that DOS applications do, it does 
not provide you with a measure of the graphics and disk performance you will 
see should you run Microsoft Windows and Windows-based applications on your 
PC. To get a picture of your PC's performance under Windows, you can use 
PC Bench in combination with two other Ziff-Davis benchmarks: WinBench and 
Winstone 94.

+       Use WinBench to measure the performance of your PC's graphics and 
	disk subsystems in a Windows environment. ZDBOp based the WinBench 
	tests on application profiling. Thus, these tests reflect the actual 
	operations applications perform. WinBench uses synthetic tests and 
	does not run applications.

+       Use Winstone 94 to measure the overall performance of your PC as it 
	runs actual Windows-based applications. With Winstone 94 you can also 
	measure the effect that changing individual system components has on 
	your PC's overall performance when running Windows and Windows-based 
	applications. Winstone 94 is the sole Microsoft-certified benchmark 
	for the testing of industry-standard PCs running Windows 3.1-based 
	applications.

You can use the form at the back of this manual to request individual copies 
of WinBench, Winstone 94, and other Ziff-Davis benchmarks from ZDBOp. You can 
also download Ziff-Davis benchmarks from ZiffNet.


What's next

Chapter 6 lists some of the factors that can influence PC Bench results and 
gives tips on how to configure your PC to get the best PC Bench results 
possible.

Chapter 7 explains the concepts behind the development of PC Bench.

If you would like to know where to find published PC Bench results, read 
Chapter 8.

If you need to run the PC Bench tests, see Chapter 3.


End of Chapter



Chapter 6
What Affects PC Bench's Results


This chapter lists some of the factors that can influence the PC Bench 
results for your PC and offers some tips on ways to get the best PC Bench 
results possible. After you read this chapter, you should understand how to 
run PC Bench to obtain optimal results.


Software and hardware factors that influence PC Bench's results

Many factors, both hardware and software, can influence PC Bench results. For 
example, CPU speed obviously affects results. Perhaps less obvious is that if 
your PC has a fragmented disk, its Disk Score will often be lower than if the 
disk is freshly formatted or defragmented.

NOTE:   There are facets of your computer's architecture you may not normally 
	consider that can affect your PC Bench results. Some PCs in the ZDBOp 
	lab showed improved performance if single SIMMs were replaced with 
	multiple SIMMs. For example, one PC ran tests faster with four 4 MB 
	SIMMs instead of one 16 MB SIMM.

The following list gives some of the software and hardware factors that can 
influence your PC's score.

Processor subsystem     CPU model and clock speed
			Internal and external CPU RAM caches
			Main memory speed
			Memory/CPU interface speed and width
			Wait states
			Power-saving settings

Disk subsystem          Hard disk drive type
			Hard disk controller and size of cache (if any)
			Hard disk driver software
			Software disk cache and size (if any)
			Fragmentation level of disk and remaining 
			   space on disk
			Compression
			Bus interface (such as local bus)

Video subsystem         Graphics adapter
			Video interface, such as Local Bus

Software                Background activity (especially network
			   and TSRs)
			Version of DOS

Hardware bus types      ISA bus, EISA bus, Micro Channel (TM) bus
			Local bus presence and type (for example, 
			   VESA or PCI)


Tips for improving your PC Bench results

The following sections provide tips on ways to get the best and most accurate 
PC Bench results for your PC.


Following proper testing procedures

The way to get the best and most accurate PC Bench results for your PC is to 
do the following each time you run PC Bench:

1.      Reformat or defragment your hard disk.

2.      Install the operating system (DOS) from scratch.

3.      Install PC Bench.

4.      Change your AUTOEXE.BAT and CONFIG.SYS files to meet PC Bench's 
	requirements.

5.      Reboot your PC.

These steps ensure that your PC is "clean" each time you run PC Bench. A 
clean configuration is a freshly rebooted PC with DOS and a minimum of other 
software.

Because reformatting your hard disk and reinstalling your operating system 
may not always be convenient or even possible, at a minimum you should always 
do the following each time you run PC Bench:

+       Defragment your hard disk.

+       Free as much RAM as reasonably possible by removing all 
	non-essential items from your PC's AUTOEXEC.BAT and CONFIG.SYS files. 
	(See Chapter 2 for the recommended AUTOEXEC.BAT and CONFIG.SYS files.)

+       Reboot your PC.

Typical PC activities, such as creating, saving, and deleting files, fragment 
your hard disk. Running the PC Bench Disk Test Suite numerous times will also 
fragment your hard disk for the same reason. Fragmentation can affect your 
PC's performance and, thus, its PC Bench results. If you run PC Bench, work 
on your PC, and then run PC Bench again without defragmenting your hard disk, 
your score may not be the same. When you defragment your hard disk, you 
return it to a relatively stable initial state. To defragment your hard disk, 
be sure to use a utility that both defragments the disk and resets the DOS 
file pointer.


Other ways to improve your PC Bench results

You may see an improvement in your PC Bench results if you:

+       Halt TSRs or network software you normally run and run only PC Bench. 
	If you do have such tasks running, they could affect your PC Bench 
	results as well as cause your PC to run out of memory during some of 
	the PC Bench tests.

+       Use disk caching.

+       Enable any processor cache(s).


Using PC Bench to best advantage

To use PC Bench to best advantage, consider how you normally use your PC. In 
most cases, the DOSMark will provide an adequate measure of your PC's overall 
performance. In addition to the DOSMark Test Suite, however, you can also run 
other individual tests according to your specific needs. For example, if in 
your day-to-day work you move lots of large files on your PC's disk, you may 
want to run the 32 MB File Disk Throughput Test.

You should run PC Bench on your PC the way you run DOS applications on your 
PC. For example, if you normally do not run applications on your PC with a 
disk cache in effect, then run PC Bench without any disk caching. The 
resulting score will reflect how you normally use your PC. If, on the other 
hand, you normally use disk caching software, use that same software when you 
run PC Bench.

Do not run PC Bench tests when you have a memory manager other than EMM386 or 
HIMEM.SYS loaded on your PC. Memory managers can interfere with the proper 
operation of protected mode tests. ZDBOp tested PC Bench 8.0 with the EMM386 
and HIMEM.SYS memory managers that ship with DOS. While PC Bench may run if 
you have a memory manager loaded on your PC, ZDBOp does not recommend using 
PC Bench as a tool for testing or comparing the performance of memory 
managers.


What's next

Chapter 7 provides a look into the concepts behind PC Bench. If you would 
like to know more about the development of the PC Bench tests, read this 
chapter.

Chapter 8 tells you where to find published PC Bench results for  use in 
comparing your PC with other PCs.

If you need to run the PC Bench tests, see Chapter 3.


End of Chapter



Chapter 7
The Concepts Behind PC Bench


This chapter discusses some of the factors that went into creating the 
PC Bench tests and making them reflect the way popular applications actually 
use PC subsystems. After reading this chapter, you should understand the key 
concepts underlying PC Bench.


The design goal

The key design goal behind PC Bench is to provide a benchmark that lets you 
accurately measure the performance of the processor, memory, video, and disk 
subsystems of a PC. The results should reflect the way real, top-selling DOS 
applications perform.

PC Bench provides such a set of synthetic performance measurement tests. Its 
results tell you why a PC responds a certain way to certain types of 
operations. PC Bench does not run actual applications. Instead, it performs 
many of the typical operations DOS applications perform.

The DOSMark gives you a single number representing the overall performance of 
your PC when running DOS applications. The DOSMark is a weighted harmonic 
mean of the scores for the Processor, Video, and Disk subsystem tests. 
PC Bench also returns scores for these three subsystem test suites. PC Bench 
weights the DOSMark and subsystem scores to make them reflect as accurately 
as possible the way popular DOS applications use PCs. (For more information 
on these tests, see Chapter 5.)

You can use the DOSMark and subsystem scores to get a general picture of your 
PC's performance. PC Bench also computes individual test results you can use 
to see more details about a particular subsystem's performance. The 
individual test results provide a way for you to explore further your PC's 
subsystem-level performance.

Some of the other factors that went into the design of PC Bench include:

+       Profiling applications to determine the most commonly used operations. 
	PC Bench follows these application profiles in the mixes it runs.

+       Using weighted harmonic means to combine groups of results based on 
	the way applications actually use PC subsystems.


Profiling applications

PC Bench uses tests based on extensive profiling of such popular DOS 
applications as the following:

+       Borland's dBASE (R) IV
+       Borland's Paradox (R)
+       WordPerfect Corporation's WordPerfect (R)
+       Lotus (R) 1-2-3 (R)
+       Microsoft (R) C
+       Harvard Graphics (R) 

Profiling is a method of monitoring an application to see what types of 
operations it performs, how often it performs these operations, and how long 
it takes each operation to complete. By using the profiling information helps 
PC Bench's tests accurately reflect what applications really do.

By following the profiles of popular DOS application, PC Bench's tests 
provide a market-centered measure of performance. The weights PC Bench uses 
to calculate its scores reflect the importance of individual operations in 
popular applications.


Understanding the "market-centered" concept

ZDBOp used a market-centered approach to develop the PC Bench tests. The DOS 
applications ZDBOp profiled were standard, best-selling applications 
performing typical user tasks. Thus, the application profiles enable 
PC Bench's tests to perform operations much like those the profiled 
applications perform. This market-centered approach lets PC Bench produce an 
accurate, overall picture of your PC's performance when running DOS 
applications.


Using a weighted harmonic mean

PC Bench generates the DOSMark, Processor, Video, and Disk scores using 
weighted harmonic means. Each harmonic mean combines several individual test 
results to create a single, representative score for a test suite. PC Bench 
places weights on the different tests to indicate the importance of an 
operation. Each weight is based on the operations actual applications 
perform. An operation an application performs frequently has a higher weight 
than an operation an application performs rarely. For example, integer CPU 
instructions are far more important than floating point calculations in the 
vast majority of DOS applications, so PC Bench weights its CPU instruction 
mixes more heavily than its Floating Point Emulation Test in calculating the 
overall Processor Score.


Understanding the Disk Mix

PC Bench 8.0 includes a new testing methodology: the Disk Mix. Today, PCs 
have both hardware and software disk cache tools available, and PC Bench's 
disk tests must accurately reflect the way applications work with these tools.

The Disk Mix performs a range of disk activities over a large set of files. 
It uses over 100 files spread across six directories. It works with multiple 
files concurrently, creates and deletes files, moves data in different sizes, 
intersperses write and read operations, and generally uses disk subsystems 
the way applications do. PC Bench's Disk Mix reflects the actual disk usage 
patterns of the profiled applications.

The Disk Mix works with a PC's disk subsystem via the standard DOS Int 21H 
interface, so you can use the Disk Score as an accurate measurement 
regardless of whether you are using a software or hardware disk cache.


PC Bench's Protected Mode Large Instruction Mix

PC Bench 8.0's 16-bit Protected Mode Large Instruction Mix reflects the 
larger working-set sizes that today's applications use. This mix accurately 
simulates the way popular applications running under DOS use a PC's processor 
and memory subsystems. The 16-bit Protected Mode Large Instruction Mix 
reflects typical application performance in such areas as bus utilization and 
use of the CPU internal cache, CPU cache interface, external cache, main 
memory, and the CPU instruction set.


Correlating the score to applications

PC Bench's DOSMark represents a cumulative effort of profiling today's 
popular DOS applications and extensive testing of PC Bench and is an 
extremely accurate measurement of overall PC performance under DOS. During 
one large test in the development process, the DOSMark yielded a correlation 
with performance tests involving profiled applications in excess of 99.5 
percent on over 70 configurations with and without SMARTDRV enabled.


What's next

Chapter 8 tells you where to find published PC Bench results for comparing 
your PC with other PCs.

For a list of all the PC Bench tests and a brief description of each, see 
Appendix A.

To run PC Bench, read Chapter 3. To view or compare PC Bench results, see 
Chapter 4.


End of Chapter



Chapter 8
How Your PC Compares


This chapter tells you where to find published results and other results to 
use in comparing PC Bench results for your PC with results for other PCs.



Checking published results
The easiest way to compare your PC to other PCs is to check articles that 
publish PC Bench results. Many Ziff-Davis publications provide PC reviews 
that include PC Bench performance scores. You can find published results in 
"First Looks{SYMBOL 228 \f "Symbol"}" articles, comparative reviews, and 
other stories in such Ziff-Davis publications as:

United States

Computer Shopper (R)
PC/Computing (R)
PC Magazine (R) 
PC Week (R) 
Windows (TM) Sources (TM)

United Kingdom

PC Direct (TM)
PC Magazine (TM)

France

PC Direct (R)
PC Expert (R)

Germany

PC Direkt (TM)
PC Professionell (TM)
Windows (TM) Activ (TM)

You can also find articles about Ziff-Davis benchmarks on Computer Library's 
Computer Select{SYMBOL 226 \f "Symbol"} CD-ROM, a database of 
computer-related product information. You may also be able to download 
PC Bench results from various forums on ZiffNet.


Using PCs within your organization
If you have access to other PCs within your organization, you can run 
PC Bench on those PCs, import their results to a single PC, and then use 
those results from within PC Bench as comparison points. (For more 
information on how to import PC Bench results, see the section "Importing 
PC Bench results" in Chapter 4.)

You can also merge PC Bench results from different PCs into one file by:

1.      Printing your current results to an ASCII text file or CSV format 
	file using the Print option from the File menu in the PC Bench 
	program screen.

	Once you have your results in ASCII text file format, you can then 
	read the text file into a database or spreadsheet program.

2.      Moving the PC Bench results for different PCs to one PC running a 
	database or spreadsheet program.

3.      Importing the ASCII text file containing the results into one 
	database or spreadsheet.

Once you have all the results in one file, you can view and print those 
results from your database or spreadsheet program.


Publishing PC Bench scores

The PC Bench License Agreement requires that you include certain information 
about your PC with any PC Bench results you publish. To find out exactly what 
you must include when you publish PC Bench results, consult the license 
agreement at the front of this manual and in the PC Bench README.TXT file.


What's next

The appendices that follow provide information you may find useful when 
running PC Bench.

+       Appendix A lists all the PC Bench tests and provides a description of 
	each test.

+       Appendix B lists error messages you may receive while running 
	PC Bench and, where possible, tells you what to do to deal with 
	these errors. This appendix also explains what to do if you must 
	abort PC Bench while it is running its tests.

+       Appendix C tells you how to get technical support for PC Bench.


End of Chapter



Appendix A
The PC Bench Tests


This appendix lists the tests that PC Bench includes and gives you a brief 
description of each. For information on how to run these tests, see 
Chapter 2.


Processor tests

The Processor tests measure the processor subsystem performance by running 
instruction mixes that reflect the way applications use a PC's processor 
subsystem. These tests exercise both the protected and real modes of the CPU, 
the CPU's internal cache (if any), and the CPU/memory interface. The 
individual processor tests are:

+       16-Bit Protected Mode Small Mix and 16-Bit Real Mode Small Mix. The 
	16-Bit Protected Mode Small Mix and 16-Bit Real Mode Small Mix tests 
	are small instruction mixes that will run within the internal cache 
	of almost all PC CPU chips that contain such a cache. These mixes let 
	PC Bench test CPU speed largely independently from the rest of the 
	PC. Thus, the results give a good, relative measure of basic CPU 
	performance.
	
	These tests use a mix of instructions based on popular DOS 
	applications' average instruction use. Both the real and protected 
	mixes reflect the instruction frequencies of DOS applications. The 
	real mode test generally should run faster than the protected mode 
	test due to the protection-checking overhead of protected mode.

+       16-Bit Protected Mode Standard Mix and 16-Bit Real Mode Standard Mix. 
	The 16-Bit Protected Mode Standard Mix and 16-Bit Real Mode Standard 
	Mix tests use medium-sized instruction mixes that should fit entirely 
	in most second-level 64 KB or larger caches. These tests use a PC's 
	CPU, CPU cache, and bus the way average DOS applications do.
	
	The instruction set mixes use the same basic technology as the 16-bit 
	small mix. These tests indicate how effectively the processor 
	interacts with the rest of the processor subsystem and how 
	efficiently a system manufacturer has interfaced the CPU to the cache 
	components. In addition, these tests serve as a good way to measure 
	the performance effect of the presence or absence of an external 
	cache subsystem.


+       16-Bit Protected Mode Large Mix. The core of PC Bench's Processor 
	Score is the 16-bit Protected Mode Large Instruction Mix. This new 
	large instruction mix reflects the larger working-set sizes that 
	today's applications use. This mix accurately simulates the way 
	leading DOS applications use a PC's processor and memory subsystems. 
	The 16-bit Protected Mode Large Instruction Mix uses a PC's bus, CPU 
	internal cache, CPU cache interface, external cache, main memory, and 
	instruction set in the same way as leading DOS applications.

+       String Sort and Move. The String Sort and Move test times how long it 
	takes the processor subsystem to execute a bubble sort on 200 random 
	strings containing 16 characters each. This test can be useful in 
	gauging the speed of systems with a RAM cache, because sorting 
	requires reading and writing some of the same data repeatedly.

+       Prime Number Sieve. The Prime Number Sieve test measures how long it 
	takes to execute a routine that finds the prime numbers between 0 and 
	8190. This is a common processor test that PC Bench includes for that 
	reason.

+       Floating Point Emulation. The Floating Point Emulation test sets up a 
	floating point emulation program in RAM and then performs a mix of 
	floating-point calculations. The processes this test performs are 
	identical to those in the Math Co-processor test.

+       Math Co-processor. The Math Co-processor test exercises the math 
	co-processor using the same floating-point calculations the Floating 
	Point Emulation test uses. You can use the test to analyze the speed 
	differences of floating point co-processors in different systems. 
	Additionally, you can use the test to gauge the processing speed 
	increase a co-processor provides by comparing a PC's Floating Point 
	Emulation results with its Math Co-processor results.


EGA/VGA text mode tests

The EGA/VGA text mode tests include both BIOS tests and direct-screen tests.

NOTE:   Some BIOS manufacturers purposely slow scrolled writes so you will 
	have time to read the scrolling text.

If the video BIOS is loaded and copied from ROM into RAM (for example, if it 
is shadowed), the throughput for these tests will be higher.

+       Unscrolled BIOS Write. The Unscrolled BIOS Write test times the speed 
	of writing data to the screen using interrupt 10h BIOS calls.

+       Scrolled BIOS Write. The Scrolled BIOS Write test times the speed of 
	writing data to the screen using interrupt 10h BIOS calls. Once the 
	test fills the screen, it scrolls through an entire screen by adding 
	one new line at a time. This test writes the same amount of data to 
	the screen as the Unscrolled BIOS Write test.

+       Direct Screen Write (8-bit) and Direct Screen Write (16-bit). The 
	Direct Screen Write tests write data to the video adapter screen 
	memory using the MOVSB (8-bit) and MOVSW(16-bit) instructions. In 
	general, one can reasonably expect the 16-bit test to yield about 
	twice as much throughput as the 8-bit test.


EGA/VGA graphics mode tests

+       Write Mode 0 Fill (8-bit), Write Mode 0 Fill (16-bit), and Write Mode 
	0 Fill (32-bit). The Write Mode 0 Fill tests fill the video adapter's 
	VGA memory with varying colors using the STOSB (8-bit), STOSW 
	(16-bit), and STOSD (32-bit) instructions with EGA/VGA write mode 0 
	and all bit planes enabled. (The STOSD instruction is only applicable 
	on PCs with at least an 80386 processor.)

+       Mem to Scr BitBlt (16-bit) and Mem to Scr BitBlt (32-bit). The Memory 
	to Screen BitBlt tests create screen segments and then save these 
	screen images to system memory one video plane at a time. The timed 
	portion of the test copies the screen segments in system memory back 
	to screen memory using the MOVSW (16-bit) and MOVSD (32-bit for 
	systems with at least an 80386 processor) instructions, one video 
	plane at a time. Very fast video cards will show the text in a 
	"marquee" effect on the screen.

+       Scr to Mem BitBlt (16-bit) and Scr to Mem BitBlt (32-bit). The Screen 
	to Memory BitBlt tests are similar to the Memory to Screen BitBlt 
	tests except that the timed portion of the test measures the 
	throughput of copying information in screen memory to system memory 
	using the MOVSW (16-bit) and MOVSD (32-bit for systems with at least 
	an 80386 processor) instructions one video plane at a time. Because 
	these tests update system memory, there is no visible activity during 
	this test.

+       Scr to Scr BitBlt (16-bit) and Scr to Scr BitBlt (32-bit). The Screen 
	to Screen BitBlt tests divide screen memory into four quadrants and 
	fill the first three quadrants with uniform fill patterns of 
	different colors. The test repeatedly moves screen data from the 
	first three screen quadrants to the fourth one using the MOVSW 
	(16-bit) and MOVSD (32-bit for systems with at least an 80386 
	processor) instructions with write mode 1.


Disk tests

The Disk Mix performs a broad range of disk activities over a large set of 
files. It uses over 100 files spread across six directories. It works with 
multiple files concurrently, creates and deletes files, moves data in 
different sizes, intersperses write and read operations, and generally uses 
the disk subsystem the way applications do.

The Disk Mix works with a PC's disk subsystem via the standard DOS Int 21H 
interface, so you can use the Disk Score as an accurate measurement 
regardless of whether you are using a software or hardware disk cache.

PC Bench's Disk Mix reflects the disk usage of a set of profiled, leading DOS 
applications. Thus, the Disk Mix works with your PC's disk subsystem the way 
actual applications do.


Disk throughput tests

The disk throughput tests show the effect different file and block size 
combinations have on raw disk performance. The disk tests use 32 MB, 16 MB, 
8 MB, 4 MB, 2 MB, 1 MB, and 256 KB file sizes to test disk throughput. For 
each file size, these tests move blocks of many different sizes: 200 bytes, 
512 bytes, 2 KB, and 4 KB. For each block size, the tests perform sequential 
reads, sequential writes, random reads, and random writes. All together, 
there are 105 disk throughput tests.

These tests are most useful for gauging three main aspects of disk 
performance:

+       The variation in DOS overhead caused by various block sizes.

+       The overhead due to the inefficiency of crossing sector boundaries. 
	The 200-byte block size causes disks to cross their 512-byte sector 
	boundaries during reads and writes.

+       The differences between in-cache and out-of-cache throughput. You can 
	get an idea of these differences by comparing the results for a file 
	size that fits in the cache against the results for a file size that 
	exceeds the cache's size.

If you want to measure your PC's disk subsystem performance when manipulating 
files with specific sizes in specific block sizes, use the Disk Throughput 
Tests. Note, however, that due to the low-level approach of these tests, they 
may not give a realistic reflection of disk performance when a cache is 
active.

The following is a list of all the Disk Throughput Tests:

32 M File       32M-200 Byte Records
		32M/200 Sequential Read
		32M/200 Sequential Write
		32M/200 Random Read
		32M/200 Random Write
	
		32M-512 Byte Records
		32M/512 Sequential Read
		32M/512 Sequential Write
		32M/512 Random Read
		32M/512 Random Write
	
		32M-2K Byte Records
		32M/2K Sequential Read
		32M/2K Sequential Write
		32M/2K Random Read
		32M/2K Random Write
	
		32M-4K Byte Records
		32M/4K Sequential Read
		32M/4K Sequential Write
		32M/4K Random Read
		32M/4K Random Write

16 M File       16M-200 Byte Records
		16M/200 Sequential Read
		16M/200 Sequential Write
		16M/200 Random Read
		16M/200 Random Write
	
		16M-512 Byte Records
		16M/512 Sequential Read
		16M/512 Sequential Write
		16M/512 Random Read
		16M/512 Random Write
	
		16M-2K Byte Records
		16M/2K Sequential Read
		16M/2K Sequential Write
		16M/2K Random Read
		16M/2K Random Write
	
		16M-4K Byte Records
		16M/4K Sequential Read
		16M/4K Sequential Write
		16M/4K Random Read
		16M/4K Random Write

8 M File        8M-200 Byte Records
		8M/200 Sequential Read
		8M/200 Sequential Write
		8M/200 Random Read
		8M/200 Random Write
	
		8M-512 Byte Records
		8M/512 Sequential Read
		8M/512 Sequential Write
		8M/512 Random Read
		8M/512 Random Write
	
		8M-2K Byte Records
		8M/2K Sequential Read
		8M/2K Sequential Write
		8M/2K Random Read
		8M/2K Random Write
	
		8M-4K Byte Records
		8M/4K Sequential Read
		8M/4K Sequential Write
		8M/4K Random Read
		8M/4K Random Write

4 M File        4M-200 Byte Records
		4M/200 Sequential Read
		4M/200 Sequential Write
		4M/200 Random Read
		4M/200 Random Write
	
		4M-512 Byte Records
		4M/512 Sequential Read
		4M/512 Sequential Write
		4M/512 Random Read
		4M/512 Random Write
	
		4M-2K Byte Records
		4M/2K Sequential Read
		4M/2K Sequential Write
		4M/2K Random Read
		4M/2K Random Write
	
		4M-4K Byte Records
		4M/4K Sequential Read
		4M/4K Sequential Write
		4M/4K Random Read
		4M/4K Random Write

2 M File        2M-200 Byte Records
		2M/200 Sequential Read
		2M/200 Sequential Write
		2M/200 Random Read
		2M/200 Random Write
	
		2M-512 Byte Records
		2M/512 Sequential Read
		2M/512 Sequential Write
		2M/512 Random Read
		2M/512 Random Write
	
		2M-2K Byte Records
		2M/2K Sequential Read
		2M/2K Sequential Write
		2M/2K Random Read
		2M/2K Random Write
	
		2M-4K Byte Records
		2M/4K Sequential Read
		2M/4K Sequential Write
		2M/4K Random Read
		2M/4K Random Write

1 M File        1M-200 Byte Records
		1M/200 Sequential Read
		1M/200 Sequential Write
		1M/200 Random Read
		1M/200 Random Write
	
		1M-512 Byte Records
		1M/512 Sequential Read
		1M/512 Sequential Write
		1M/512 Random Read
		1M/512 Random Write
	
		1M-2K Byte Records
		1M/2K Sequential Read
		1M/2K Sequential Write
		1M/2K Random Read
		1M/2K Random Write
	
		1M-4K Byte Records
		1M/4K Sequential Read
		1M/4K Sequential Write
		1M/4K Random Read
		1M/4K Random Write

256 K File      256K-200 Byte Records
		256K/200 Sequential Read
		256K/200 Sequential Write
		256K/200 Random Read
		256K/200 Random Write
	
		256K-512 Byte Records
		256K/512 Sequential Read
		256K/512 Sequential Write
		256K/512 Random Read
		256K/512 Random Write
	
		256K-2K Byte Records
		256K/2K Sequential Read
		256K/2K Sequential Write
		256K/2K Random Read
		256K/2K Random Write
	
		256K-4K Byte Records
		256K/4K Sequential Read
		256K/4K Sequential Write
		256K/4K Random Read
		256K/4K Random Write


DOS disk access

The DOS disk access group of tests includes the BIOS Disk Seek (sequential 
and random) and the DOS Disk Access tests.

NOTE:   If a seek error occurs in either of the tests, PC Bench will display 
	a message for a few seconds, choose a replacement sector, and then 
	continue the test. Such a failure is not necessarily an indication of 
	a bad disk drive. These tests access sectors that may have been 
	legitimately locked out during the formatting of the drive. If you 
	suspect a problem with your drive, use a diagnostic program to test 
	the drive.

+       BIOS Disk Seek - Sequential and BIOS Disk Seek - Random. The BIOS 
	Disk Seek tests (sequential and random) measure mechanical 
	track-to-track disk drive access times. The tests use the standard 
	BIOS interrupt 13h interface. You cannot use this test to check 
	floppy diskettes or mass storage media that do not have an Interrupt 
	13h interface. The tests uses 1,000 sequential seeks (alternating 
	between cylinders 0 and 11) and 1,000 random seeks. Caching disk 
	controllers and certain SCSI adapters that do not force the disk 
	mechanism to move for disk seeks may yield unrealistically fast 
	times.

+       DOS Disk Access. The DOS Disk Access Test measures the time necessary 
	to perform 1,000 read requests at random locations on the disk using 
	the DOS Int 25H interface. This test should generally work with any 
	device DOS recognizes as a disk except CD-ROM and network drives. 
	The test reports the total time (in seconds) required to complete the 
	requests.


Memory timing tests

The memory timing tests gauge a PC's speed at accessing various types of 
memory. A PC's CPU speed and the architecture of its memory, memory caches 
(if present), and CPU can affect the results of these tests.

The memory timing tests include a set of Extended Memory tests. These tests 
enter protected mode to access memory above 1 MB. As a result, these tests 
will generally not execute properly or fail to execute at all if an extended 
memory driver (such as EMS, DPMI, or VCPI) is already installed on your PC. 
The memory tests basically use the full range of extended memory available. 
The write tests avoid any area already locked by an XMS (HIMEM) driver, RAM 
disk, or other DOS Int 15H-compatible software.

There are six Extended Memory tests, a read test and a write test for each 
data width. Those tests are:

+       8-bit Extended Memory Read and Write
+       16-bit Extended Memory Read and Write
+       32-bit Extended Memory Read and Write

The memory timing tests also include a set of Graphics Memory tests. These 
tests detect the type of VGA graphics adapter present and use the appropriate 
buffer addresses to read and write blocks of data directly to the graphics 
adapter's VGA memory. During memory writes, PC Bench will overwrite the 
screen's contents and then restore them at the end of the test.

There are six Graphics Memory tests, a read and a write for each data width. 
These tests are:

+       Graphics Memory 8-bit Read and Write
+       Graphics Memory 16-bit Read and Write
+       Graphics Memory 32-bit Read and Write

The memory timing tests also include BIOS Memory tests. These tests read 
only from the BIOS. These tests generally shows significant performance 
differences when your PC's BIOS ROM is shadowed or cached, a common 
configuration choice on many systems. There are three BIOS memory tests:

+       BIOS Memory 8-bit Read
+       BIOS Memory 16-bit Read
+       BIOS Memory 32-bit Read

The memory timing tests also include a set of Text Memory tests. These tests 
detect the type of VGA display adapter and use the appropriate screen buffer 
addresses to write directly to the display adapter's memory.

There are four Text Memory tests, a read and a write test for each data 
width. These tests are:

+       Text Memory 8-bit Read and Write
+       Text Memory 16-bit Read and Write

Finally, the memory timing tests also include a set of Conventional Memory 
tests. These tests access memory in the 640 KB area that DOS manages.

There are four Conventional Memory tests, a read and a write test for each 
data width. These tests are:

+       Conventional Memory 8-bit Read and Write
+       Conventional Memory 16-bit Read and Write


Tests included as utilities

PC Bench includes the following tests you can run as utilities. You execute 
these tests from the directory where you installed PC Bench.


Battery Rundown Test

The Battery Rundown Test exercises a portable computer's battery system in a 
worst-case scenario. You should start the test after fully charging the 
computer's battery according to the manufacturer's specifications and 
disabling all power-conservation features. The test types 10 screens full of 
information to the computer's display and then writes a 15-KB file to disk 
with a time stamp of elapsed time. PC Bench repeats this operation until the 
battery runs out of power.


Time the Timer Test

The Time the Timer Test calculates elapsed time using DOS calls. Some systems 
may not report a correct time when running a version of DOS that is not 
properly configured for the system. In this case, the results for the Time 
the Timer Test may be skewed. This test lets you measure the accuracy of the 
system's timer by comparing it to some external source (such as a stopwatch).


VGA Compatibility Test

The VGA Compatibility Test performs several VGA register-level functions that 
exercise the different subsystems (including the VGA controller chip, video 
memory, and RAMDAC) of a VGA or VGA-compatible display subsystem. Because the 
test is an older test designed with the original VGA in IBM PS/2 systems in 
mind, most boards do not pass every test but can still run application 
software without any problems. Thus, a failure of a few of the tests does not 
usually indicate serious compatibility problems for most software.


EGA/VGA Monitor Quality Test

The EGA/VGA Monitor Quality Test displays images in various modes from 
black-and-white text to VGA color graphics. You can use these tests to 
determine what display modes a monitor or display adapter combination can 
support. Also, you can use these tests to evaluate monochrome portable 
computer screens to see how well they map colors to gray scales.


End of Appendix



Appendix B
Troubleshooting


Normally, PC Bench runs without problems. Occasionally, however, you may 
encounter a problem when you run PC Bench. This appendix gives you some basic 
information about what to do if you have problems with PC Bench. It also 
lists messages you may receive when running PC Bench and tells you what to do 
when you receive those messages.

If you have a problem and cannot resolve it after reading this appendix, 
write down any message (including any directory or file names) that appears 
on your PC's screen, note what PC Bench was doing at the time you received 
the message (if you know), and then contact ZDBOp's PC Bench technical 
support as described in Appendix C of this manual.


What to do if you have problems

PC Bench may encounter a fatal error if it runs out of memory, or it may stop 
running (hang) during its tests on some PCs due to compatibility issues. In 
most cases, PC Bench will exit as gracefully as it can if there is a serious 
problem. If you have problems, you should follow a few basic steps before you 
contact technical support:

+       Verify that your PC meets the hardware and software requirements of 
	PC Bench. Also, check to be sure that you have set up your PC 
	correctly. (For more information on both areas, see Chapter 2.)

+       If PC Bench issues the error "Unable to allocate enough memory," you 
	will need to free more conventional memory by removing TSRs or device 
	drivers or by loading DOS high with HIMEM.SYS. (For more information, 
	see the section "Setting up your PC to run PC Bench" in Chapter 2.)

+       If you get a PC Bench database error or a message suggesting you 
	restore your database files, you should exit PC Bench. You should 
	then restore PC Bench's database files by typing the following 
	command from the directory where you installed PC Bench:

	RESTDB

	This command restores the backup copy of your database files and may 
	fix any corruption that may have occurred. The backup files are 
	usually those that were present the last time you started PC Bench.

	NOTE:   Even though you restore PC Bench's database file, you will 
		lose any data that you added to the database since the last 
		backup.

+       If PC Bench seems to hang during a test or some operation and is not 
	responding to any input, you may need to reboot your PC. To reboot, 
	you can:

	*       Press Ctrl-Alt-Del

	*       Press your PC's reset button if it has one

	*       Turn the PC's power off

	NOTE:   Several PC Bench tests may take a while to run. One way to 
		tell if PC Bench is still running its tests is to check to 
		see if your PC's disk light is active. If it is, then 
		PC Bench is still running its test. Do not abort PC Bench 
		unless you are very sure there is a problem.

	Restart PC Bench and see if the problem persists. If so, you can 
	contact PC Bench technical support at ZDBOp.


Running PC Bench with memory managers

PC Bench 8.0 works with the EMM386 and HIMEM.SYS memory managers that ship 
with DOS.

Although PC Bench may run if you have a different memory manager loaded on 
your PC, ZDBOp does not guarantee your results. Also, ZDBOp does not 
recommend using PC Bench as a tool for testing or comparing the performance 
of memory managers.

NOTE:   If you are running the Above DISK (R) memory manager (either by 
	running PC Bench on a 286 CPU or running Above DISK on a 386 or 
	higher CPU), PC Bench may hang or may not execute its tests 
	correctly. PC Bench requires a 386 or higher CPU to successfully run 
	its tests.


Running PC Bench in batch mode

If you have problems running PC Bench from a DOS batch file, try running it 
with the /M option. This option prevents PC Bench from logging your PC's 
configuration information. You should not stop PC Bench from logging your 
PC's configuration information unless you are having problems running 
PC Bench.

You should also note that using the /P option when running PC Bench in batch 
made can cause problems on future PCs with new processors or on DOS emulators
. This option prevents PC Bench from automatically determining what type of 
processor is in your PC.


PC Bench error messages

This section lists most of the error messages you might receive while running 
PC Bench and explains what you should do if you get one of these messages.

   You may need to free up XX bytes on your system before PC Bench can 
   successfully run all its tests at once.

	PC Bench issues this warning message if you try to run PC Bench with 
	less than 540 KB of free conventional memory. You need a minimum of 
	540 KB of free conventional memory to run all the PC Bench tests at 
	once.

	When you get this error, you can continue if you do not plan to run 
	all the tests at once, or you can exit PC Bench, free conventional 
	memory by removing TSRs, drivers, etc., and then restart PC Bench. 
	Remember, if you change your CONFIG.SYS or AUTOEXEC.BAT files, you 
	must reboot your PC before you restart PC Bench.

   PC Bench supports only VGA (640x480) resolution or greater. PC Bench may 
   run on your display, but text in the PC Bench screens will be truncated.

	ZDBOp designed PC Bench to run on VGA systems. If you run PC Bench on 
	a non-VGA system, it will try to continue, but the display and 
	positioning of its windows and buttons will not be correct. 
	Generally, you should not run PC Bench without VGA support.

   PC Bench does not support processors less than 80386. If you continue, you 
   will not be able to run all the PC Bench tests.

	You get this error message if you attempt to run PC Bench on a PC 
	with a CPU less than an 80386 (or 80386-compatible). PC Bench 8.0 
	requires a minimum of a 386 (or compatible) CPU.

   PC Bench can only produce a DOSMark score and the Protected Mode Large Mix 
   Processor results for PCs with more than 1 Mb of Extended Memory.

	You will receive this error message if PC Bench detects that there is 
	not enough extended memory on your PC to run the Protected Mode Large 
	Mix. This test requires just over 1 MB of extended memory. PC Bench 
	is only able to calculate the DOSMark score on PCs with a total of at 
	least 2 MB of RAM.

   Unable to allocate enough memory.

	This message indicates that PC Bench has completely run out of 
	conventional memory and cannot continue its current operation. 
	PC Bench will quit after it displays this message. You can exit 
	PC Bench and free up conventional memory by removing TSRs, drivers, 
	etc. Even if you do not change your CONFIG.SYS or AUTOEXEC.BAT files, 
	you should reboot your PC before you restart PC Bench to ensure the 
	tests run correctly.

	Note that memory fragmentation can also cause low memory conditions. 
	Try repeating the operation that caused this error message after you 
	restart PC Bench before doing any other operations.

   Insufficient memory to run test.

	Not enough conventional memory is available to provide the currently 
	running test with the buffers it needs to run. The test will cancel, 
	and if there is another test still to run, that test will begin. That 
	test, however, may also cause the same error. This error message is 
	not fatal and does not keep PC Bench from continuing.

   Insufficient memory to load test group.

	This message is similar to the one above, but it indicates that there 
	was not enough conventional memory to hold the test code itself. In 
	this case, PC Bench skips the whole test group (for example, 
	processor tests) you are trying to run. If you get this error 
	message, you need to free more conventional memory to run the 
	PC Bench tests.

   This machine's settings are not unique. Please change its Machine ID or 
   Variants.

	You will get this error when you are using the Machine Information 
	dialog box if you set the Machine ID and Variants to the same 
	settings as an existing set of results. You should change one or more 
	of these fields.

   The Machine ID and Variants you specified are not unique.

	You will get this error in batch mode when you specify the DOS 
	environment variables Machine ID or Variants that are not unique. You 
	should re-enter these settings.

   The Machine ID and Variants are not unique for the new machine.

	You will receive this message when PC Bench tries to automatically 
	create variants for your PC but none of them are unique. This might 
	happen if you name a series of PCs with variants that just happen to 
	coincide with the current date and time. If you wish to use the 
	variants that conflict with the current date and time, you should try 
	again after some time has passed, or you can change the Variants 
	using batch mode.

   Invalid file name.
	
	You receive this fatal PC Bench error if you use the /Ffile option 
	and file is not a valid file name. You should verify that the file 
	name you supplied is correct.

   Could not find the environment variables:
     PIN
     VARIANT1
     VARIANT2
     VARIANT3
     VARIANT4
     VARIANT5

	You get this message when you use the /B option without setting the 
	minimum necessary DOS environment variables. To correct the problem, 
	make sure you set the appropriate variables when you execute PC Bench 
	from the DOS command line or from a batch file.

   You cannot execute this program directly--Use BENCH.

	You will receive this message if you try to run a PC Bench test 
	module from the DOS command line. You cannot run these modules 
	individually; they will run only when PC Bench starts them.

   Internal Error--Please contact ZDBOp.

	If you receive this fatal PC Bench error, it means PC Bench has 
	encountered an internal error. Write down what PC Bench was doing at 
	the time you received the error, if possible. Then contact ZDBOp as 
	described in Appendix C.

   The PC Bench licensing information is not set up correctly. You should 
   re-license PC Bench.

	Before you can run PC Bench, you must first license it. You will 
	receive one of these error messages if PC Bench cannot find the 
	license file it creates when you license and register the program. To 
	license PC Bench, use the /L option when you start PC Bench.

   The licensing procedure failed.

	You receive this error message when the PC Bench licensing procedure 
	aborts (possibly due to a file error). Try licensing PC Bench again 
	with the /L option.

   PC Bench cannot find a resource file. Check to be sure you are in the 
   PC Bench directory.

	You may receive this error message if you attempt to run PC Bench and 
	you are not in the directory where you installed it. Check to make 
	sure you are in the correct directory.

   Open Error: FILENAME.DBF

	You may receive this message when PC Bench encounters an error trying 
	to open a database file. FILENAME is the name of the database file. 
	Your database of results may be corrupt. You should restore your 
	PC Bench database files and restart PC Bench. If you still get this 
	error, contact ZDBOp.


General database errors

When PC Bench encounters general database errors, it will automatically issue 
a warning message and quit. You should reboot your PC and restore PC Bench's 
database files.

NOTE:   When you restore PC Bench's database files, you will lose any data 
	you have added to the database since PC Bench backed up its files.

General database errors are of the format:

	Unique Key error
	(Error XX)
	Restore your database files

Where XX is an internal error number.

You can  record this number and contact ZDBOp.


End of Appendix



Appendix C
Technical Support


This appendix tells you how to get copies of PC Bench and other Ziff-Davis 
benchmarks and tells you how to order printed copies of this manual. This 
appendix also explains the different ways you can contact the Ziff-Davis 
Benchmark Operation (ZDBOp).

Requesting a benchmark

PC Bench is available free of charge on 3.5" diskettes. You can get copies of 
PC Bench and other Ziff-Davis benchmarks free of charge by contacting ZDBOp 
or by downloading the benchmark you want from ZiffNet.

NOTE:   The Winstone 94 CD-ROM includes both PC Bench and WinBench. Thus if 
	you get the Winstone 94 CD-ROM, you actually get all three benchmarks.

To request a benchmark from ZDBOp, use the form that comes after this 
appendix. You can fax the form to the ZDBOp dedicated benchmark request fax 
number (919-380-2879), or you can mail the form to ZDBOp at the address 
listed at the top of the form. The benchmark will arrive via third-class US 
mail in four to six weeks.

If you want to receive a benchmark sooner and you have a Federal Express 
account, include your Federal Express account number and shipping 
instructions on the fax form. The benchmark will arrive via Federal Express 
and ZDBOp will charge the cost of sending the benchmark to your Federal 
Express account.


Ordering a printed manual

You can also order a perfect-bound, printed copy of this manual for a nominal 
fee. To order a copy of this manual or other Ziff-Davis benchmarks' manuals, 
use the Ziff-Davis Benchmarks' Manual Order Form included at the back of this 
manual.


Contacting ZDBOp

If you have comments or questions about PC Bench, please contact ZDBOp. You 
can contact ZDBOp in several ways:

+       If you have a modem and communications software, you can reach ZDBOp 
	via ZiffNet, the Ziff-Davis on-line service. One way to access 
	ZiffNet is by using CompuServe (R). ZiffNet's 
	ZDBOp forum (GO ZDBENCH) is a way for you to communicate your 
	technical questions and comments to Ziff-Davis. (Access to CompuServe 
	is available for a fee.)

+       You can fax questions and comments directly to PC Bench Technical 
	Support at the ZDBOp fax number (919-380-2879).

+       You can mail questions and comments to ZDBOp at the following 
	address:

	Ziff-Davis Benchmark Operation
	One Copley Parkway, Suite 510
	Morrisville, NC 27560
	
	Attn: PC Bench Technical Support


End of Appendix


Benchmark Request Form

FAX OR MAIL THIS FORM TO:

Ziff-Davis Benchmark Operation
One Copley Parkway, Suite 510
Morrisville, NC 27560

fax: (919) 380-2879

WHO I AM:

Name: ____________________________________________________________
Company: _________________________________________________________
Address: __________________________________________________________
City: _________________________ State: _____________ Zip: ______________
Country: __________________________________________________________
Fax: ___________________________ Phone: _____________________________

We answer requests in the order we receive them. We ship all benchmarks via 
third-class US mail unless you supply a Federal Express account number. 
Please allow four to six weeks for delivery via regular mail.

Shipping information if you choose to use Federal Express:

Your Federal Express account number: ______________________________
Check one: ___ priority overnight ___standard overnight

WHAT I WANT:

___ PC Bench (TM) 8.0           ___ NetBench (TM) 2.1
___ WinBench (R) 4.0            ___ ServerBench (TM) 1.0 SCO (R) UNIX (R)*
___ Winstone (TM) 94 Version 1.0 (on CD-ROM)
___ MacBench (TM) 1.0

(All benchmarks EXCEPT Winstone 94 are available only on 3.5" diskettes.)

* Other versions to follow



Ziff-Davis Benchmarks' Manual Order Form

To Order:

Fax this form to (919) 380-2879 or mail to:

Ziff-Davis Benchmark Operation
One Copley Parkway, Suite 510
Morrisville, NC 27560

Attn.: Manual Orders

Title                                           Price   Qty.    Total
Understanding and Using WinBench (R) 4.0        $10.00  ____    _______
Understanding and Using Winstone (TM) 94 
	Version 1.0                             $10.00  ____    _______
Understanding and Using PC Bench (TM) 8.0       $10.00  ____    _______
Understanding and Using MacBench (TM) 1.0       $10.00  ____    _______
Understanding and Using NetBench (TM) 2.1       $10.00  ____    _______
Understanding and Using ServerBench (TM) 1.0 
	for SCO (R) UNIX (R)                    $10.00  ____    _______

Shipping and handling included.                 TOTAL   ____    _______

Ship to: (please print)

Name:   ________________________________________________________________
Company:        _____________________________________________________________
Address:______________________________________________________________
City / State / Zip:     ________________________________________________________
Daytime Phone:  _________________________________________________________

Payment form:

__Check to Ziff-Davis Publishing Co.    __Visa  __MasterCard    __American Express
Credit Card #: _____________________________________      Exp. Date: ___________

Signature: ______________________________________________________________

We do not accept Purchase Orders or cash payments.

For fastest shipment supply your Federal Express or UPS account number:

Your account number:_____________________________________________________
Check one:      __ Priority overnight   __Standard overnight



Acknowledgments

The Ziff-Davis Benchmark Operation (ZDBOp) developed version 8.0 of PC Bench. 
Numerous people worked together to create this product. Members of the 
primary PC Bench development team are:

	Debbie Willmschen       PC Bench Technical Documentation
	Thomas E. Siering       PC Bench Development Leader
	Michael Agerbak         PC Bench Developer
	David J. Arendash       PC Bench Developer
	Bradley G. Epranian     PC Bench Developer

In addition, the PC Bench benchmark would not be possible without the 
use of software from Phar Lap Software, Inc.

Other contributors to the contents of the PC Bench diskette include:

	Richard Butner
	Gabriel DeBacker
	Israel Ehrisman
	Dan Ghidali
	Homi Kapadia
	Kasey Lee

The ZDBOp support staff who helped make PC Bench possible are:

	Betsy Bates
	Jennie Faries
	Rae Fernandez
	Gina Massel-Castater
	Olivia Ongpin
	Maria Rowan

Many people in different parts of Ziff contributed to the design, testing, 
and production of PC Bench, including:

	Tim Baltrusch
	Adam Baum
	Jessica Beyer
	Michael Blakeley
	Brain Booher
	Dean Bushik
	Stephen Chan
	Michael Edelhart
	Robert Ferguson
	Jim Galley
	Edward O. Griffin
	Edward Henning
	Alex Ho
	Bert Jensen
	Bob Kane
	Amy Leung
	William Machrone
	Cynthia Meade
	Jim Nell
	Huy Nguyen
	Ron Panus
	Harold Poskanzer
	Dave Salvator
	Kern Schireson
	Nick Stam
	Tim Stefanini
	Jim Wolf
	Joe Virzi












