NBM - The Ness Benchmark
v1.2
A) What it is...
NBM is a program which runs on any Atari ST/STe/TT computer, in any
resolution. NBM runs five tests, timing their duration, for use in
comparing the speed of the machine with other Atari computers.
NBM includes a Compare function, which contains a database of
previously tested Atari computers of various configurations. This
allows the user to compare results with other people's machines.
Also included is a Print function, allowing the user to print the
results of each test.
B) Disclaimers...
NBM gives specific results for specific tests. The tests performed
may or may not portray anything a user does in day-to-day operation of
the computer. As a result, it should be noted that some of the tests
may not be significant to you, the user.
On the other hand, the five tests included in NBM do attempt to
portray legitimate routines a computer will be doing in as many
different real applications as possible.
If nothing else, NBM allows the user to compare the advantages
and/or disadvantages of various hardware and software combinations.
NBM has been tested on hundreds of machines of various
configurations, with no reported ill effects. However, the author
accepts no responsibility for any damages, real or perceived, resulting
from the use of this software. Use NBM at your own risk.
C) How to use it...
Load the program as you would any other GEM graphics program. NBM
displays a dialog box which lists the six tests to be performed. Also
displayed are your TOS operating system version, and whether or not the
program is running in fast (TT) or normal (ST) ram.
Across the bottom are buttons labelled Test, Print, Compare and
Exit. Before tests are performed, the Print and Compare buttons are
disabled.
The test sequence is begun by clicking on the Test button, but you
must first select which of the six tests are to be performed. This is
done by clicking on the test descriptions themselves. When selected for
testing, each selection will have a checkmark in front of it. Clicking
again, on a selected test, will deselect it.
As each test is being run, its description will appear in reverse
video. As it is completed, its results will appear on the right side of
the dialog box.
After all selected tests have been run, the Print and Compare
buttons at the bottom of the box become active. Choosing Print will
cause the results to be printed out. Your Atari Control Panel chooses
printing parameters. Make sure your printer is turned on.
Choosing Compare will run you through a database of previously
tested machines. Your results for a particular test are displayed at
the top, while other results appear in order from slowest to fastest.
D) The results...
NBM results are displayed in two columns. The first column
indicates how much time, in seconds, each test took to complete. The
second column shows a percentage comparison between the tested machine,
and a "baseline" machine. The baseline machine is a TOS 102 ST, running
at 8mhz, with no Desk Accessories or AUTO programs installed.
E) The tests...
In the Math test, NBM performs 1000 iterations of a combination of
integer and floating point math. The floating point math does not take
advantage of any math coprocessor which may be installed.
The Memory test performs 25 iterations of a very large block memory
move. Approximately 250,000 bytes of ram are moved from one location to
another, and back again, in ram.
In the Dialog Box test, a standard GEM dialog box, with many items
in it, is drawn 150 times, making heavy use of the graphics routines
built into GEM.
The next two tests involve disk drives. The Floppy Drive test
accesses drive A: on your machine, and attempts to open and manipulate a
250,000 byte file.
The Other Drive test accesses whatever drive NBM resides on, and
performs 4 iterations of the same test done on your floppy drive. This
test is useful in analyzing the speed of your hard drive or
ramdisk.
The Graphics test runs your ST through three sub-tests - text, line
and object. The text test repeatedly prints a line of text, until the
window is filled, then changes attributes and starts again. The line
test draws lines from one side of the window to the other, from about a
dozen starting points. The object test draws a number of objects, some
rectangular with square edges, some rectangular with round edges, and
some oval.
The Graphics test time will vary depending on both your machine's
speed, and the size of the screen you are using. The test takes place
within a window sized to almost fill your screen, and the various tests
are designed to use the full window.
F) Notes...
There are many things which can affect the results you get on your
machine.
First and foremost, DO NOT MOVE THE MOUSE during a test. The
internal routines which sense mouse movement and draw the mouse onscreen
use system resources. While those routines are working, the test is not
running. Your times will be misleadingly long. Simply click on the
Test button and take your hand off the mouse.
Second, every Desk Accessory, and some AUTO programs, use system
resources while NBM is running. Properly written, the effect is small,
but they will make a noticeable difference in your test results. One of
the interesting things that NBM can do is to show you exactly how much
your system is affected by these programs.
In the dialog box and graphics tests, your results will vary if
you have one of the popular software screen accelerators installed
(QuickST, TurboST, NVDI). NBM is a good medium from which to make a
gross comparison of these products.
The disk drive test results will vary wildly, based on a number of
factors. Disk fragmentation is the most important. A disk which
contains a number of files, especially files which are regularly written
to, can easily become fragmented. This will slow down any disk drive
test dramatically. The machines displayed in the Compare dialog did not
have fragmented drives.
In the case of hard drives, different brands and models will show
slightly different results. This is simply because hard drive speed is
a sales feature. Manufacturers are always trying to make their drives
faster.
For ramdisks, the ramdisk software is the important issue. Some
software is more efficient than other software. This affects how
quickly data is stored to and retrieved from the ramdisk.
G) Talk to me...
I would be very happy to hear from you regarding your experiences
with NBM. I can be contacted on CompuServe (74415,1727), GEnie (J.NESS)
and Delphi (JNESS).
NBM is a FREEWARE product. Use it to your heart's content, don't
worry about having to pay for it. But, if you pass it around, you must
distribute it without modification, and include this documentation file
intact with it.