Contents of the SPEED600.DOC file
S y s t e m S p e e d T e s t
V e r s i o n 6.00
Copyright (C) 1993 Landmark Research International Corporation
All Rights Reserved
Valid command-line switches for this program are as follows:
/b - bypass this initial screen
/nv - don't perform initial video test
/q - run in quiet mode (no sound)
/xx - run for 'xx' seconds, then exit
You may control this program with a mouse and/or keyboard.
To scroll the text in this box, click on the up and down arrows in the text
box border, or press the up/down arrows, or page up/down keys.
To change the scale of the speed bars (normally yellow), click the left or
right mouse button while the mouse cursor is on top of the bar. Left button
increases the scale to make the bar shorter (violet), and right button
decreases the scale to make the bar longer (cyan, or light blue). Press F3
and F4 to increase/decrease the CPU and FPU scales, or F5 and F6 to change
the video scale.
If you are running on a monochrome system, you will not be able to discern
what CPU type or speed scales are active from a distance because color is
used for those purposes.
Press F1 for help, F9 to recalculate video speed, F10 to toggle the beeper
on and off, and Esc to exit the program.
You may run the program with one or more command line parameters separated
by spaces to affect operation:
SPEED /B - bypass initial help screen
SPEED /NV - no video (disable video test)
SPEED /Q - quiet (disable beeper)
SPEED /40 - run for 40 seconds and exit
SPEED /B /NV /Q /40 - all of the above
The system speed tests tell you the most important information about the
computing performance of your system: the effective speeds of the central
processor unit (CPU), floating point unit (FPU, math coprocessor),
and video display circuitry. Although hard disk access and data transfer
rate affect system performance, we excluded those tests because they do not
affect actual computation speed.
The CPU and FPU speeds show performance relative to an IBM AT with a 6 MHz
80287 math coprocessor chip installed. For example, if the display shows
CPU speed to be 22.53 MHz, it means an IBM AT would have to run at 22.53
MHz to have the same computing performance as this computer. If the display
shows FPU speed to be 40.25 MHz, it means an IBM AT would have to run
its 80287 math chip at 40.25 MHz to have the same floating point math
calculation performance. The speeds are NOT the actual CPU/FPU clock speeds.
The CPU test forces the processor to show the effect of instruction cache.
It is thus more accurate than previous versions. On color systems, the
background display is gold for 8088 and 8086 CPUs, green for 80286,
blue for 80386, and red for 80486. This allows you to see from a distance
what kind of CPU is running the system. The display also shows in text the
type of CPU installed.
The FPU test shows a value only if a math chip is installed or built into
the CPU. Where possible, it shows the manufacturer of the chip (e.g., Intel,
Cyrix, IIT). Note that the 80486 has an internal FPU, but other CPUs do not.
The test accurately reflects the higher performance of a closely coupled
The program runs both CPU and FPU tests continually, and updates the display
about 4 times per second. The speed will drop if you run another application
concurrently or change CPU clock speed. You will be able to see the system
speed degradation caused by multitasking or background network accessing,
or by taking the sys
tem in and out of turbo mode.
Clock Speed / Multitasking
Multitasking does not affect actual CPU clock speed, but allows less time
for the speed test to run; turbo mode actually changes the CPU clock speed,
and thus affects the performance of both multitasking and the speed test.
Some systems do not have a turbo mode, but those that do usually allow you
to invoke it with a front panel switch or the keyboard (normally
Ctrl-Alt-plus or Ctrl-Alt-minus).
As the test runs, you may notice the CPU and FPU MHz values changing
slightly. This does not mean your system clock is running erratically,
but rather that test results are being rounded off, and occasional
interrupts in the background are using CPU time.
If you want the fastest possible readings, you should boot DOS on the
computer with no drivers or resident programs installed. Such software can
cause the CPU to work on special tasks in the background while the speed
test runs in the foreground. The special tasks will take CPU time and cause
the speed test to give lower readings than it should.
Large temperature changes can cause PCs to vary in performance
characteristics. As a result your computer may run faster when it is cold
than when it is warmed up.
The video test checks pure data transfer rate from the CPU into video memory.
It thus shows the maximum speed at which data can be read and displayed.
The video test detects most popular video adapter chipset types, and displays
the type and manufacturer on the screen. If it cannot determine the
manufacturer, it shows just the generic type, such as monochrome, CGA, EGA,
Video / System Design
The video test results are affected by video adapter design, CPU speed, and
system board design. For example, some systems have fast I/O bus speeds
and allow a fast data transfers to video adapters. Others, even some with
very fast CPU speeds, have slow I/O speeds and thus slower display rates.
Much of the power of a fast video adapter can be wasted if it is used in a
slow system. On the other hand, a slow video adapter can impede the faster
systems, giving diminished performance in display-intensive applications.
Video Runs Once
The program runs the video speed test only once, at the beginning. This is
partly for aesthetic reasons, since it makes the display look unattractive.
You may re-run the video test by pressing F9 (recalculate). You will need
to do this if you take the system in or out of turbo mode while the normal
speed test is running, because video speed can be affected by CPU speed.
The video test does not show the effect of BIOS on display speed because
most popular high-performance applications write directly to video memory,
and do not use BIOS. Furthermore, most super VGA adapters come with a
program that loads display BIOS into system RAM for maximum execution speed.
All of the test results are shown on variable scales. The default scale
values are different for each CPU type, but the default speed bar is always
yellow. This allows the program to show how the performance of your
system compares to the scale limits for your processor type.
Thus, if you view the screen from a distance, the program accurately renders
performance information without the exaggerated effect of display wrap-around
or logarithmic scaling used in earlier versions of the program. If you see a
yellow speed bar, you are seeing the default scaling for your CPU type.
If you are running on a computer that is slow for the class of CPU it uses,
the speed bars will be short. If the computer is fast, the speed bars will
be longer. If the bar is too short or too long for your liking, you may
change its length by adjusting the scale.
However, when you change the scale, you will also change the color of the
speed bars to either cyan or violet. This will allow distant viewers to
recognize that the system is not using the default scale and that the
length of the bars have a different meaning.
If the bar is too long on a scale, you have a very fast computer. To shorten
it and turn the bar violet, press F3 or click the left mouse button while the
cursor is on top of the bar. F3 expands CPU and FPU scales, while F5 expands
the video scale.
If the bar is too short on a scale, you are running on a relatively slow
computer. To lengthen the bar and turn it cyan, press F4 or click the right
mouse button while the cursor is on top of the bar. F4 shrinks CPU and FPU
scales, while F6 shrinks the video scale.
HISTORY AND STANDARDS
From 1984 through early 1993 the original Landmark CPU Speed Test (tm)
gradually gained prominence and became the de facto world standard for
measuring PC computing speed. There are several reasons for its rise to
1. Standard - it compared the speed of the computer it ran on to the speed
of the first and only de facto standards that ever existed for personal
computers: the IBM PC and the IBM AT. IBM never changed the architecture
of either computer, but replaced them with the PS/2 line in 1987 and took
the PC and AT out of production. Since IBM will never change the PC or AT,
those will always be stable standards of comparison for PC performance,
and Landmark will be the performance measurement standard. The original PC
used an Intel 8088 CPU and optional 8087 FPU running at 4.77 MHz. The
original AT used an Intel 80286 CPU running at 6 MHz, and a later version
ran at 8 MHz; in both versions the optional 80287 math chip ran at 6 MHz.
2. Visibility - the Landmark speed bar that reached horizontally from left
to right across the screen was visible and relatively readable from a
distance of 15 to 30 meters (50 to 100 feet). This made it ideal for
dealer showrooms and trade shows, so manufacturers and resellers enjoyed
using it to show off their computer's speed.
3. Credibility - it was the most accurate benchmark test on the market
because it considered a variety of factors that other programs did not, and
it showed performance relative to the PC and AT. Other benchmark programs
gave unreliable or misleading information:
Norton System Index (SI) was too simple and overrated AT performance relative
to the PC.
MIPS (millions of instructions per second) used an instruction mix that was
Whetstone, a test of general computing combined with floating point or
scientific math calculations, was originally written in FORTRAN, so the
results were badly skewed by both compiler efficiency and the ratio of
floating point math instructions to arithmetic and loop instructions.
Dhrystone, a test of general computing performance, was originally developed
to test ADA compiler efficiency, but was converted to the C language for
execution on the PC. Because of the plethora of C compilers in use, it
suffered from the same kind of skewing as Whetstone.
4. Results - The speed test displayed its results graphically with fixed
scaling, and it beeped while it gave continuous screen updates. All the
other programs, such as SI, MIPS, Power Meter, CheckIt, and QA Plus, ran
through once and stopped. Thus, in real time it showed how much system
performance degraded during network accesses or other multitasking program
activities. Furthermore, the instant update gave immediate feedback to
users, and gratified them that the system was actively running. Finally,
benchmark graphic displays in QA Plus and CheckIt automatically scaled test
results so that their speed bars had no meaning when viewed from a distance.
Landmark Research International Corporation is a high-technology research,
development, manufacturing, marketing, and sales organization with
headquarters in Clearwater, Florida.
The original company, Landmark Software, was founded in 1984 in Sunnyvale,
California, to produce the Landmark CPU Speed Test, Landmark Setup program,
and assorted system utility programs. In early 1990, the company was
restructured with expanded business goals under new leadership and given a
new base of international operations on Florida's west coast.
Landmark develops and supplies computer-related products to many types of
customers: dealers, distributors, computer manufacturers, 3rd and 4th party
service companies, corporate micro management and repair centers, and
Existing products include the following:
1. The PC Probe (tm) advanced testing program for PCs - it features replete
system information, complete diagnostics, a full suite of benchmark tests
based on the Landmark System Speed Test, CMOS RAM setup and low-level hard
disk format utilities, remote operation, and the ability to display
actual photographs of the circuit or component being tested. The complete
and perfect program for end users, technicians, dealers, and OEMs to find
out what's in the computer and how well it works.
2. The AlignIt (tm) maintenance kit for 3.5 and 5.25-inch floppy drives of
all popular densities - it includes a pre-lubricated cleaning diskette
guaranteed for 180 10-second uses, and a patented diagnostic and alignment
spiral diskette that tests and allows drive alignment to within .3 mils.
Perfect for technicians and end users alike to avoid costly drive repairs.
3. The KickStart (tm) family of diagnostic cards for PCs - the cards assist
technicians in identifying failures in systems that won't boot, and help
Value-Added Resellers to run remote diagnostics on client's computers.
LEDs or digital displays show failing power-on self-test numbers during boot.
Built-in circuitry on some models provides serial, parallel, floppy, and
hard-disk I/O, as well as special test hardware for motherboards. The design
engineer's and test/repair technician's constant companion.
4. The Burn (tm) advanced manufacturer's burn-in test for PCs - it features
menu or command line operation via local console or remote terminal, nested
count-based or time-based testing loops, pre-configured user-modifiable test
sequences, and error logging to printer, file, or terminal. Site licenses
available to PC system and motherboard manufacturers.
5. The Memory Boss (tm) family of Expanded Memory Managers - fully
compatible with Lotus-Intel-Microsoft Expanded Memory Specification 4.0,
Memory Boss operates on XTs (disk-based emulator), ATs (extended
memory based emulator), 386 ATs (virtual 86 mode based emulator), and
Chips and Technologies NEAT chipset EMS circuitry. Gives Lotus, Windows,
and other applications huge memory space to operate in. Perfect for
all levels of users; OEM licenses available.
6. The JumpStart (tm) family of self-configuring motherboard BIOS for 80286,
386, and 486 ISA and EISA computers as well as XTs. Fully IBM-compatible and
non-infringing. Available in ROM for system developers and troubleshooters.
Source and object code (label) licenses available for manufacturers.
Other products that have been in development and may be available include
graphic image processing and file conversion; FAX/Data modems with voice mail
software; automated test equipment cards for ATs; and software developer
libraries for XT/AT/386 AT virtual memory management, graphic image data base
management, and menu systems. Contact Landmark for more information.
When you think of standards, think LANDMARK.
The Author cannot and does not warrant that any functions contained in the
Software will meet your requirements, or that its operations will be error
free. The entire risk as to the Software performance or quality, or both,
is solely with the user and not the Author. You assume responsibility for
the selection of the program to achieve your intended results, and for the
installation, use, and results obtained from the Software.
The Author makes no warranty, either implied or expressed, including with-
out limitation any warranty with respect to this Software documented here,
its quality, performance, or fitness for a particular purpose. In no event
shall the Author be liable to you for damages, whether direct or indirect,
incidental, special, or consequential arising out the use of or any defect
in the Software, even if the Author has been advised of the possibility
of such damages, or for any claim by any other party. But if you print it
out, light it on fire and your house burns down, don't blame me.
All other warranties of any kind, either express or implied, including but
not limited to the implied warranties of merchantability and fitness for a
particular purpose, are expressly excluded.